|
Edit cuppatea did some nice numbers about tournaments, watch it before looking at the graphic.
I used some numbers from (sc2ranks.com) to look at the Racial Distribution at high niveau. My visualisation of the numbers can be found here (a version from two days later):
![[image loading]](http://img839.imageshack.us/img839/9784/racediamond.png)
While it is known that Blizzard's Matchmaking Algorithm is functional (TL Thread), it was not clear how the same win percentage per race was computed. Some told that there is a racial divergence between diverse skill levels (Bronze League to Diamond League), because new player tend to stay at the campaign's race Terran and the Zerg mechanics may be harder to learn. An alternative to this hypothesis is that people tend to chose race's which let them easier win. If this would be true we should see a race more often if it makes winning simplier. This does not mean that if a race wins often this race would be more easy to play.
Some suggested to compare the racial distribution (RD) at high skill levels as indicator for imbalance or more likely as indicator for the lack of exploitable gamestyle. I think this indicator should be seen as a trend on higher skill levels - and more important as it's change, if either new strategies and tactics are discovered, or there are gameplay changes (patches, etc.) implemented.
The upcomming patch will change gameplay in special situations quite heavily and we should see a shift in the high level racial distribution over some time (player's Inertia) till an equilibrium is found again (I am not sure if there will be an equilibrium, but it could exist). The time of the shift and the endtime will be good situations to overlook some presumptions we can create right now.
I think it would be interesting to do estimations about the post patch use of various units, since it gives us a little time to figure out new concepts before they are established on all servers. This thread could be general discussion about those changes - if it gets more strategy biased it could be moved to the strategy directory.
*Edit: Added Sample Count*
*Edit: Good post, with things I would like to see happen*
*Edit: Data as CSV*
*Edit: Added the "plus" to the 1500 Intervall*
*Edit: Someone made a picture of the absolute numbers as you see the comparision is impossible without adjusting the scale*
*Edit mahnini did a nice post with a combined intervall 1100 to 1400*
*Edit Cotonou did a post in which he tries to motivate a null-distribution with the percentage of gold's league (P/T/Z/R 38%/32%/20%/10%) - if he is right, we should see no shift at all on the high level after the patch is released, but the same which happens on the gold level.*
*Edit: Good arguments:
The percentage will change fast in the top 20, my argument: we can look at the trend of the top 200. There is a racial bias in every league and we have to look at distribution's change to tell stuff. There is a temporal error in it which can be cleared with a second measurement right after the patch gets uploaded.
But still my question is: What are we able to tell with this data? What would we think will happen if the patch changes gameplay? Is there a distribution's equilibrium right now? This are the important questions I really would like to see answered.
*
*Edit: Added the version from today * + Show Spoiler +
*Edit: Toxigen made a nice comment and linked us back to the beta's history and the old racial distribution.*
*Edit: Scarmath crawled sc2ranks.com and grouped players into bins, before finally performing Individual Chi Square tests and shows us his findings.*
*Edit Scarmath and Sentinent have done some calculation I agree with and visualized them on this page ff.*
|
am i reading this wrong... or is it a lot easier to be a low level diamond player as terran than any other race?
which i firmly believe... haha =/. A Zerg and Terran player of equal skill... the Terran will get more wins/points/easier division bump... ?
just my 2 cents... ya i guess im crying imba =/
|
This is so LOOOL oh dear good please NERF all terran units
|
On September 02 2010 07:43 mnofstl007 wrote:
am i reading this wrong... or is it a lot easier to be a low level diamond player as terran than any other race?
which i firmly believe... haha =/. A Zerg and Terran player of equal skill... the Terran will get more wins/points/easier division bump... ?
just my 2 cents... ya i guess im crying imba =/
You're reading it wrong. Protoss has the most from 600-1100, and then Terran takes over from there.
|
it means absolutely nothing because the sample size is tiny at the right end of the graph
|
I'm going to guess that around 1200 is when people are able to stop 4 gate pretty well.
|
8748 Posts
On September 02 2010 07:46 cup of joe wrote:
it means absolutely nothing because the sample size is tiny at the right end of the graph
There is no sample size. This is a graph of the entire population.
|
On September 02 2010 07:46 cup of joe wrote:
it means absolutely nothing because the sample size is tiny at the right end of the graph
It appears to be 28k players. That is rather large compared to the total population. Even if it was, assuming the distribution is normal within diamond, even 1000 players would be more than enough to make reasonable conclusions with a reasonable margin of error.
|
Thats pretty convincing, sure the sample size is smaller at the top end of the graph but your talking about very refined players at that area as well. This means less inconsistency in the results of the game ( ie losing to dumb mistakes, off racing, or cheese loses/wins ). We're taking about basically the top 600 players in the world where it is extremely terran heavy ( I checked on sc2 rankings, the 600th ranked person is very close to the 1200 mark ).
|
'high niveau'? Random French...
This is so LOOOL oh dear good please NERF all terran units
yeah let's not. I thought that this was what people mocked WoW players for.
|
On September 02 2010 07:46 cup of joe wrote:
it means absolutely nothing because the sample size is tiny at the right end of the graph
I disagree that it means absolutely nothing. Because regardless of numbers, the right end of the graph represents the "top players", and it shows that amoung them, Terran is dominate.
However, it would be nice to see the sample size there to put it into context.
|
On September 02 2010 07:45 Dionyseus wrote:Show nested quote +On September 02 2010 07:43 mnofstl007 wrote:
am i reading this wrong... or is it a lot easier to be a low level diamond player as terran than any other race?
which i firmly believe... haha =/. A Zerg and Terran player of equal skill... the Terran will get more wins/points/easier division bump... ?
just my 2 cents... ya i guess im crying imba =/ You're reading it wrong. Protoss has the most from 600-1100, and then Terran takes over from there.
ooooooo
i read it backwards.
its a lot easier for terran to be at the TOP of their diamond leagues... got it thx
|
On September 02 2010 07:50 Liquid`Tyler wrote:Show nested quote +On September 02 2010 07:46 cup of joe wrote:
it means absolutely nothing because the sample size is tiny at the right end of the graph There is no sample size. This is a graph of the entire population.
Not quite. He means that there are so few players at the highest point levels that you can't use that information reliably.
|
I guess something about Terran just naturally attracts all the really sick players, right? (/sarcasm)
But it's interesting how the % of Terran takes off like an exponential while zerg/protoss fall off. Random is perhaps the most interesting; it looks like some really sick players enjoy playing it (maybe for the challenge?), while not-as-sick players try it but cant succeed to the degree that they might if they picked a race and mastered it.
The random peak in zerg at 1300 suggests cup of joe could be correct in asserting that the sample size is too small at these higher point values. Otherwise the trends seem pretty clear.
EDIT: As of Wednesday night, 20 players are over 1500, 48 between 1400 and 1500, 92 from 1300 to 1400, and 205 between 1200 and 1300, for a total of 365 people in the range that shows terran dominance.
|
8748 Posts
As far as I know, sc2ranks is pretty damn comprehensive of Diamond, especially high Diamond (where it seems some people have a problem with the "sample size"). Yeah, the number of people in the 1500+ group is small but that doesn't mean that there is a problem with the sample size. These numbers aren't extrapolated from a small population of the 1500+ Diamond group. These numbers directly represent that group.
|
it means absolutely nothing because the sample size is tiny at the right end of the graph
Almost certainly a terran player who doesnt want your godmode race to be nerfed...
The sample size is significant enough to give good results and for whatever reason it is showing that terran players make up 60% of 1500+ players. It would be very interesting to see this in comparison to the amound on players overall.
For example if 60% of all players play terrain and 60% of players at 1500 diamond are terran its nothing unexpected. However if 33% of players are terrain and 60% of players at 1500 diamond are terran then there obviously is some cause for concern
|
On September 02 2010 07:52 Sentenal wrote:
I disagree that it means absolutely nothing. Because regardless of numbers, the right end of the graph represents the "top players", and it shows that amoung them, Terran is dominate.
However, it would be nice to see the sample size there to put it into context.
You can't just say 'regardless of numbers'. If there was 1 player and he was Zerg there would be 100% Zerg.
~15 people isn't completely meaningless but it's not too huge either. It's hard to tell whether actual gameplay concerns or just perception are more influential. Either way, we shouldn't just go crazy on a kneejerk reaction. We need small steps, and also to factor in maps and such.
|
@cup of joe even if the sample size is small on the right side it still means something... You just have to be more carefull about what you are telling. For example the 1500 entry got a total of 20 which is big enough for my opinion. If you go to 1400 you even got 36 and a little more down at 1000 you got 971 entries. If you watch some geographical studies you often ask 40 to 200 people, leading to a subset of a dozen people which is quite bad, but often enough to motivate further questioning.
In fact I think that the top 100 player are quite more important regarding the game's ultimate balance than the player below that. I guess that the top 100 player on the ladder aren't the top 100 player in the world, though.
We need small steps, and also to factor in maps and such.
@Redmark I absolutely agree, I think it is clear that minor gameplay changes can have huge impacts (nerfing the reaper speed down by 0,1 means one extra hit off creep). I think the map factor is quite important if we watch the distribution's change. Korean BW Pro-League maps are much more special than most of the SC2 maps, right now.
|
Pretty eye-opening, it seems.
I'm glad that this graph pulls up an interesting point though, Terrans aren't really that overpowered except at high levels in the hands of 1000+ point diamond players.
|
So, as long as you are not 1200+ diamond, imbalance is too small to have an effect? I can live with that, thought it is worse. What about win percentage, does it favor terran in top diamond that much as well? *is checking sc2ranks.com*
|
8748 Posts
On September 02 2010 07:53 Vexx wrote:Show nested quote +On September 02 2010 07:50 Liquid`Tyler wrote:On September 02 2010 07:46 cup of joe wrote:
it means absolutely nothing because the sample size is tiny at the right end of the graph There is no sample size. This is a graph of the entire population. Not quite. He means that there are so few players at the highest point levels that you can't use that information reliably.
Use it reliably for what? He says it's not possible to answer a question using this data without indicating what the question is...?
There is nothing wrong with the data and the right end of the graph does not rely on a sample.
|
Looking at the data, anything past 1400, the data sample is too small (5z/22t/10p for 1401-1500 and it gets smaller from there).
However, it is pretty convincing that Terran seems to be dominating at very high level of play. Terran needs a nerf? That's what patches 1.1 and on are for.
|
8748 Posts
On September 02 2010 07:59 EliteAzn wrote:
Looking at the data, anything past 1400, the data sample is too small (5z/22t/10p for 1401-1500 and it gets smaller from there).
Too small for what?
|
Popularity does not equal strength.
The fact that people have this *idea* that terran might be OP and that zerg might suck will make them play terran more. You see terran all the time at high levels simply because there are more terran than zerg. The more people cry "terran OP," the more people will switch over to it.
The top players are there because they are the best players. Not because of their race. The reason there are so many terrans at the top is because there are so many terrans overall.
|
On September 02 2010 07:52 Sentenal wrote:Show nested quote +On September 02 2010 07:46 cup of joe wrote:
it means absolutely nothing because the sample size is tiny at the right end of the graph I disagree that it means absolutely nothing. Because regardless of numbers, the right end of the graph represents the "top players", and it shows that amoung them, Terran is dominate. However, it would be nice to see the sample size there to put it into context.
There's currently 65k players in the diamond league, only 27k diamond players are at least 600 points.
source: http://www.sc2ranks.com/stats/league/all/1/all
|
So on lower level, with Protoss, its easier to win ( proxy 2 gate, 4gate all in), but on top tier, protoss looks balanced, and zerg underpreforming compared to terran
Thats pretty much sums up every thread full of whines, that was made the last 4 weeks.
|
this graph means jack shit when it doesn't even indicate how many people are within each grouping.
ex. what if only 10 people were in the 1500 range
|
That is a bit disturbing.
Looking at the 600-1100 range all I can think of is 4 gate. It would be nice to see the number of people in each group between 1200 and 1500 if there are very few people it could be said that racial imbalance is only affecting the highest levels of play severely. If there is a decent number of people between that range then it could suggest that the reason there are fewer terran players between 600 and 1100 is because most of the diamond worthy terran players are in the higher ranges.
It would be interesting to see what is up with the random players, it follows a strange curve.
|
I got a Theory a little off topic:
I am a Daimond ranom player, the higher I get, the less I spawn as Terran, the more as Zerg.
Is blizzard using us to compansate the lack of race distribution? :O
---Na, just kidding, but out of 20 games today I spawned friggin 17 times Zerg xD.
|
Nice graph, shows clearly that Zerg dwindle at the very top. I don't understand why so many people responding to this thread can't read a simple graph.
|
On September 02 2010 08:01 Liquid`Tyler wrote:Show nested quote +On September 02 2010 07:59 EliteAzn wrote:
Looking at the data, anything past 1400, the data sample is too small (5z/22t/10p for 1401-1500 and it gets smaller from there). Too small for what?
Too small to make a very solid point that Terran is OP and Blizzard screwed up big time. For example (this is an extreme case), I can say Zerg is Extremely OP at the Extremely high Bronze level (There is 1 zerg/player who has 2001+ points in Bronze (lol)).
Yes, it's a trend and we can all see that the # of zerg players is decreasing as points increase. Same said with Terran but the opposite (increasing)
|
isnt it obvious?
1200-1500 terran gets higher and higher in number, which shows that the race if used right is too strong
|
|
|
i'm not too sure what to make of this graph. it appears protoss is the most common race and evenly distributed. also it appears terran players have an easier time reaching the top of the ladder than are other races although maybe just maybe it's because there are more of them?
|
Zerg is consistently the lowest, which may be a result of less Zerg players, but right between mid and high ELO, there is a shift between Protoss being the top and Terran being the top. This may be a clear indication that Terran's strength increases much more rapidly as the player's skill increases compared to Protoss.
|
On September 02 2010 08:07 EliteAzn wrote:Show nested quote +On September 02 2010 08:01 Liquid`Tyler wrote:On September 02 2010 07:59 EliteAzn wrote:
Looking at the data, anything past 1400, the data sample is too small (5z/22t/10p for 1401-1500 and it gets smaller from there). Too small for what? Too small to make a very solid point that Terran is OP and Blizzard screwed up big time. For example (this is an extreme case), I can say Zerg is Extremely OP at the Extremely high Bronze level (There is 1 zerg/player who has 2001+ points in Bronze (lol)). Yes, it's a trend and we can all see that the # of zerg players is decreasing as points increase. Same said with Terran but the opposite (increasing)
how can it be too small? it are the best player lol
if it would be bigger then there would be way to many not good player inclueded
look at broodwar u have flash jeadong and nobody more at the top
|
At the same time, doesn't this mean the terran imbalance is only really determinative at the 1000+ diamond level (although it may be easier for terrans to compete in general, it seems that its not unbeatable until it reaches this level)
|
On September 02 2010 07:55 Liquid`Tyler wrote:
As far as I know, sc2ranks is pretty damn comprehensive of Diamond, especially high Diamond (where it seems some people have a problem with the "sample size"). Yeah, the number of people in the 1500+ group is small but that doesn't mean that there is a problem with the sample size. These numbers aren't extrapolated from a small population of the 1500+ Diamond group. These numbers directly represent that group.
This post states perfectly the idea that everyone seems to be missing when complaining about sample size. I hope people read it and attempt to understand it.
|
On September 02 2010 08:01 Liquid`Tyler wrote:Show nested quote +On September 02 2010 07:59 EliteAzn wrote:
Looking at the data, anything past 1400, the data sample is too small (5z/22t/10p for 1401-1500 and it gets smaller from there). Too small for what?
LOL these posts are too funny,
there is an obvious trend im sure having 1000 extra people in the 1400++ range wouldn't all off a sudden give zerg an upward spike
|
Honestly, I hope people don't read too much into these numbers. There can be a lot of factors, and seeing it as a function of time would be important, not just a snapshot.
Everyone is still (as far as I can tell) still evolving strategies, it might be that zerg is harder to master. It is very interesting though, almost reminds me of SC1 with protoss being popular in the D/C range.
|
On September 02 2010 08:01 Backpack wrote:
Popularity does not equal strength.
The fact that people have this *idea* that terran might be OP and that zerg might suck will make them play terran more. You see terran all the time at high levels simply because there are more terran than zerg. The more people cry "terran OP," the more people will switch over to it.
The top players are there because they are the best players. Not because of their race. The reason there are so many terrans at the top is because there are so many terrans overall.
Your post is completely misinformed, and while this logic might hold true for the lower ranges (where terran is actually NOT overly popular, protoss is the most popular. Your logic is actually very flawed when concerning top level players
And the bold sentence is completely and utterly fallacious.
|
What i see from this graph is that at lower levels -before 1200- (meaning not with a PERFECT control over the game) people tend to win more with Protoss . After 1200(pros are here) lets say all players are of the same skill ( none is better than the other ) which must mean terran has some advantage with increase in macro/micro . Hope i made my point clear
id schtrudel
code 678
|
ITT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected.
"Terran players make up a disproportionate amount of 1300+ Diamond ladder players in a way that can't be attributed to random chance." - Fact? You can't readily see this from the data. You must use something like a chi-square test to determine the confidence you can attach to this statement. I think it would be a good thing if someone did this with the raw data. Preferably someone with a more sophisticated background than one year of college statistics (moi).
"Terran players make up a disproportionate amount of 1300+ Diamond ladder players and it is because Terran is imbalanced." Fact??? This would be very hard to determine conclusively, and you can't even ascribe a confidence to this claim with the data available. However, under reasonable assumptions, it is a reasonable claim.
Other claims, like "Terran is more prevalent in higher leagues because more people play it because of the campaign" are more readily testable, but require sampling of campaign play data with respect to skill level.
|
I don't think people get it.
The higher the rating, the more weight it has as variables such as player skill is minimal since they are all pros. Nobody cares about statistics in the lower levels, there are way too many variables to account for.
|
You also have to keep in mind there are a lot more protoss players in general and a lot less zerg.
A little off topic but damn DeMuslim is just eating away the competition, is he really that good that he can have a win percentage of 83.3% and still be at the top of standings? Maybe he has an awesome build against each race with little counters? Maybe he started late and plays during off hours also?
|
United States10774 Posts
On September 02 2010 08:10 Cade wrote:Show nested quote +On September 02 2010 07:55 Liquid`Tyler wrote:
As far as I know, sc2ranks is pretty damn comprehensive of Diamond, especially high Diamond (where it seems some people have a problem with the "sample size"). Yeah, the number of people in the 1500+ group is small but that doesn't mean that there is a problem with the sample size. These numbers aren't extrapolated from a small population of the 1500+ Diamond group. These numbers directly represent that group. This post states perfectly the idea that everyone seems to be missing when complaining about sample size. I hope people read it and attempt to understand it.
lol yeah, this is correct. people don't seem to understand what "sample" size means
|
The sample size is not too small for the top end, to be sure someone can just do a t-test and see if it's statistically significant that ~50 something % of the top 340 players are Terran, I'm pretty sure that'd be pretty significant. 60% for the top 20 players is also pretty significant, someone should do a t-test or a chi-square
|
On September 02 2010 08:01 Backpack wrote:
Popularity does not equal strength.
The fact that people have this *idea* that terran might be OP and that zerg might suck will make them play terran more. You see terran all the time at high levels simply because there are more terran than zerg. The more people cry "terran OP," the more people will switch over to it.
The top players are there because they are the best players. Not because of their race. The reason there are so many terrans at the top is because there are so many terrans overall.
Uh huh, you just keep thinking that. Meanwhile every tournament except MLG in the past 2 months are being won by Terrans. In reality, there aren't THAT many race switchers. I can't think of any known SC players who have switched over to Terran, but still they keep winning everything.
|
The upcoming patch is a joke. A JOKE. Zerg needs more toys, more viable strats. They could use another spell caster. Or a speed upgrade for Hydras.
None of those would make the game imbalanced.
|
On September 02 2010 08:20 Gigaudas wrote:
The upcoming patch is a joke. A JOKE. Zerg needs more toys, more viable strats. They could use another spell caster. Or a speed upgrade for Hydras.
None of those would make the game imbalanced.
Or they just need to learn how to play. (dont ban me)
|
I don't know how anyone can read this as anything but evidence that Terran is OP. Automatic match making keeps everyone at around %50 win ratio everywhere but the very top. The only way to tell balance with AMM around is by looking at the very top, and that's what this graph does. Terran is way overrepresented at the top, hence it is OP.
And I can't agree with people saying that the graph only shows imbalance at the top so it's fine everywhere. As has been said many times the top is the only place it *can* be showed with AMM in the mix. And it's there.
And for all the zerg players lower in diamond, this graph doesn't show that it's balanced for them at their level. They might be at a higher level (increasing their representation on the right side of tthe graph) if they weren't underpowered. Same for zerg in silver. They are getting %50 wins thanks to AMM but some of the higher ones might be in gold if it were more balanced.
With an even distribution of skill this isn't going to show until you look at the very top. This graph does, and the result is there plain as day.
|
Not quite. He means that there are so few players at the highest point levels that you can't use that information reliably.
But why not?
If being a top end player means you have the top ends points, you are taking into consideration 100% of the top end population. Sample size is irrelevant. for example if the average top player (major tournament caliber) player has 1400+ points, then by taking into account that group you have looked at 100% of the top player population.
people saying it should balance for all levels of play dont understand that its like saying the rules/balance of a sport at the professional level should take into consideration the weekend warrior who goes to play football on sundays.
edit: misquote
|
being a 1100+ zerg player this makes me sad in the pants. Honestly these numbers do go to show that a trend is emerging among top level players... This isn't really big news though. The longer a patch isn't out the more those numbers will grow for terran and decrease for zerg and slightly for toss.
|
if you include 1600+
then Protoss is the highest, with 3 protoss, 2 terran and 1 zerg being in the 1600's.
also note, the top 5 go zerg protoss terran protoss protoss terran, so i really dont think theres any imbalance other than how you can go semifast reaper and only be 1-2 scv's behind your opponent and have an awesome, often endgaming tool to harrass with. which is getting nerfed soon. so quit whining everyone terran is not OP.
|
On September 02 2010 08:08 Doomrok wrote:Terran is perfectly balanced, I mean, look at my Terran vs Zerg win rate! http://www.danrok.com/stats/ r r:r
Always said that the best way to evaluate balance is to look at the win rates of the different matchups of the random players. Removes the "the terrans are simply better players" argument.
Any idea how to get this kind of data for all randoms on b.net?
|
On September 02 2010 08:01 Backpack wrote:
Popularity does not equal strength.
The fact that people have this *idea* that terran might be OP and that zerg might suck will make them play terran more. You see terran all the time at high levels simply because there are more terran than zerg. The more people cry "terran OP," the more people will switch over to it.
The top players are there because they are the best players. Not because of their race. The reason there are so many terrans at the top is because there are so many terrans overall.
If this was true, then wouldn't you expect more Terran at every level, rather than steadily increasing amounts of Terran up to the very top? Or are you saying that the better the player, the chances that he just happens to enjoy Terran more increases?
Yes, there are lurking variables, for things like people-sticking-to-the-campaign-race or as you mentioned switching-to-Terran-out-of-fear. However, I wouldn't expect those variables to have such a large effect, and again, it would be even per level of Diamond, not steadily increasing.
I think Terran is simply easier to play, and because this is a game that requires a great deal of concentration/apm/macro/micro, a race that's easier will also be better, and thus imbalanced.
On September 02 2010 08:14 blacktoss wrote:
ITT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected.
"Terran players make up a disproportionate amount of 1300+ Diamond ladder players in a way that can't be attributed to random chance." - Fact? You can't readily see this from the data. You must use something like a chi-square test to determine the confidence you can attach to this statement. I think it would be a good thing if someone did this with the raw data. Preferably someone with a more sophisticated background than one year of college statistics (moi).
"Terran players make up a disproportionate amount of 1300+ Diamond ladder players and it is because Terran is imbalanced." Fact??? This would be very hard to determine conclusively, and you can't even ascribe a confidence to this claim with the data available. However, under reasonable assumptions, it is a reasonable claim.
Other claims, like "Terran is more prevalent in higher leagues because more people play it because of the campaign" are more readily testable, but require sampling of campaign play data with respect to skill level.
Thank you for this post.
P.S. to any zerg players out there, this might be amusing... I'm a bronze leaguer, and at my level of play, Terran suck!!! Ever since I switched my main to Zerg, I've beat every Terran I've met
|
Just shows that on the very top , terran > toss and zerg. And people saying its just more terran players thats why they are more on top... Then why is it suddently increasing so much after 1100 points?
That cant be because the best players all play terran...
I saw game between madfrog(z) and fuzer(t) today on go4sc2. Fuzer went banshee harass that failed first time. Madfrog had allmost won the game right there, but fuzer got planetery fortress expand with few towers around that was impossible to break without heavy loss. From there he got couple cloaked banshee into the widly spread expos of madfrog, and killed 30+ drones. Then a massed thor+hellion army destroyed the ultra/ling/muta army of madfrog, not even close.
This was on fuzers stream it it looked like he had like 70 apm, and wasnt even macroing, letting his energy on orbital command get to 100 etc.
Couple games later fuzer meets morrow, and morrow makes him look like a total newbie.
Another example was a terran that made it to the quarter finals in todays go4sc2. He went all in every game with mass reaper into reaper/marauder and sent all his scv, EVERY single game. That got him to the quarter finals...
These people who cant see that terran are very very strong now, from all the evidence provided in numberous of threads are the bad terran players, who still get raped.
|
On September 02 2010 07:46 cup of joe wrote:
it means absolutely nothing because the sample size is tiny at the right end of the graph
thats like saying that oil CEO's are not really wealthy because sample size is too small.
|
Dominican Republic463 Posts
You guys still dont get that theres no sample, its stats from the whole population? Stop saying sample size. ITT Terrans coming to the rescue to deny their race is OP, they've been shown stats, analysis, everything. but nope, we just still need to l2p!
|
On September 02 2010 08:01 Backpack wrote:
Popularity does not equal strength.
The fact that people have this *idea* that terran might be OP and that zerg might suck will make them play terran more. You see terran all the time at high levels simply because there are more terran than zerg. The more people cry "terran OP," the more people will switch over to it.
The top players are there because they are the best players. Not because of their race. The reason there are so many terrans at the top is because there are so many terrans overall.
Actually only 30% of diamond players are Terran, the most popular race for diamond players is Protoss with 35%. Zerg is at 24% and Random is at 10%.
Source: http://www.sc2ranks.com/stats/league/all/1/all
|
It means that at the highest level of play, there are more terrans than all other races combined.
If you are looking at the whole population, for instance, let's say that population is "Highest level of play", then there is no need to speak of sample size... you're looking at the whole population, there is no estimate.
Is this totally indicative that there is imbalance? No. But it supports that idea.
- If many experts are saying terran is imba, it again supports it.
- If blizzard says terran is imba and will tweak in 1.1, it again supports it.
- If experts in other races do not refute that terran is imba, it again supports it.
So, saying that at the top level of competitive play, terran is imbalanced is supported by:
- Data above.
- Experts opinion (Idra, Sheth, Morrow, etc. etc.).
- Designers opinion (Blizzard Patch 1.1 status update).
Whether it's true or not, I'm not really vested in or against... as I dont play starcraft 2, but these three sources all point toward the same idea and that gives the idea weight.
|
On September 02 2010 07:55 Liquid`Tyler wrote:
As far as I know, sc2ranks is pretty damn comprehensive of Diamond, especially high Diamond (where it seems some people have a problem with the "sample size"). Yeah, the number of people in the 1500+ group is small but that doesn't mean that there is a problem with the sample size. These numbers aren't extrapolated from a small population of the 1500+ Diamond group. These numbers directly represent that group.
You would think people would realize this. It makes me cringe when they cry about sample size when ITS THE ENTIRE POPULATION.
|
I hate graphs. People always misread them and use them to prove their ideas when they really have no idea how to read said graph.
|
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200.
|
Lol at the misinformed "sample size" people. What matters is the TREND. Look at the trend: as we increase in points in the diamond league, the number of Zerg players dwindle and the number of Terran players increase.
|
For those who think the Korean server is different.... (sry no pretty graph)
901-1000: 73z/112t/109p
1001-1100: 39/66/68
1101-1200: 19/30/40
1201-1300: 8/23/14
1301-1400: 9/11/12
1401+ : 4/9/2
|
this is highly misleading, as there are only very few people who are in 1500 region, its unfair to do a % comparison. It can vary quite drastically, which is shown in your graphs. This applies to 1200 and up. There are simply too few players to do a % comparison.
|
On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200.
We have a winner. Someone who actually knows the most basic stats calculations while other people ramble on about "sample sizes".
|
On September 02 2010 08:35 biology]major wrote:
this is highly misleading, as there are only very few people who are in 1500 region, its unfair to do a % comparison. It can vary quite drastically, which is shown in your graphs. This applies to 1200 and up. There are simply too few players to do a % comparison.
There are 360 people in the world 1200 and up. This is plenty to do statistical analysis. See my post above.
|
On September 02 2010 08:32 Wr3k wrote:Show nested quote +On September 02 2010 07:55 Liquid`Tyler wrote:
As far as I know, sc2ranks is pretty damn comprehensive of Diamond, especially high Diamond (where it seems some people have a problem with the "sample size"). Yeah, the number of people in the 1500+ group is small but that doesn't mean that there is a problem with the sample size. These numbers aren't extrapolated from a small population of the 1500+ Diamond group. These numbers directly represent that group. You would think people would realize this. It makes me cringe when they cry about sample size when ITS THE ENTIRE POPULATION.
so lets say im inventing a new pill to cure something, and i test it on a handful of ppl, lets say 5.. then those 5 ppl i gave my pill to make up the whole "population" the pill was ever tested on. so lets say the test showed no adverse reaction for any of the 5 ppl. given that information, would u like to go ahead and try my new pill?
btw: t is slightly imbalanced. they are going to fix it.
|
Maybe a suggestion, can we have numbers instead of percentages? Then less people will be complaining about how small the "sample" is....and by sample, i mean complete, raw, data.
|
On September 02 2010 08:32 Alou wrote:
I hate graphs. People always misread them and use them to prove their ideas when they really have no idea how to read said graph.
The problem is, people on the internet will attempt to use ANYTHING to prove their ideas, and will attempt to disprove ANYTHING that denies their ideas. As this thread is an obvious example of. People refuse to be scientific and accept statistical evidence, and instead fight to the death for whatever they WANT to be correct.
|
so lets say im inventing a new pill to cure something, and i test it on a handful of ppl, lets say 5.. then those 5 ppl i gave my pill to make up the whole "population" the pill was ever tested on. so lets say the test showed no adverse reaction for any of the 5 ppl. given that information, would u like to go ahead and try my new pill?
btw: t is slightly imbalanced. they are going to fix it.
And those five people would be the "sample". Here the people above 1200 are not the sample, they are the entire population of people above 1200.
|
Run a chi square first to see if this is disproportionate. I really don't have much time... but if I could get this data I could run it:
% of Tarran at x points and above (maybe 1000?)
% of Tarran for the whole population
I can deduce from that what the % is for the rest. A chi square would tell you how likely this distribution would occur randomly in the population, and something tells me the top has a disproportionate amount.
If the chi square determines it to be less than an alpha of 5% (there is less than a 5% chance of this distribution to occur), then we can concur that there is a real imbalance. What attributes to this imbalance is not given by the test. It would just say one exists.
Now, the common response would be "but everyone still has close to a 5% win rate!". This is true and this is actually what would cause this. The match making system places you in the spot where you are most likely to get a 50% win rate. If you do better, then you are placed higher. If the Tarrans doing better, the system would just place them higher so they would still have a 50% win rate. The end result would be Tarrans being over represented in the higher end of the population.
Remember, this takes into account that Protoss and Zerg are not played as much. This asked the question "are these two the same population?", meaning, do Tarrans make up the same percent at the top as they do for the rest taking into account random variance. Therefore, any argument from "oh, but less play Zerg and Protoss" does not even matter because this measures the difference between the amount of players (in percent) between the populations, not the amount itself.
Just get me the data, PM me when it's here, and I'll run the test. Someone else can do it with me to double check.
|
Hmmm really makes you think. I wonder whats going to happen after patch 1.1, will the reaper and tank nerf really make a difference?
Hopefully more zerg throughout the diamond league by the end of September.
|
On September 02 2010 08:39 StarDrive wrote:Show nested quote +
so lets say im inventing a new pill to cure something, and i test it on a handful of ppl, lets say 5.. then those 5 ppl i gave my pill to make up the whole "population" the pill was ever tested on. so lets say the test showed no adverse reaction for any of the 5 ppl. given that information, would u like to go ahead and try my new pill?
btw: t is slightly imbalanced. they are going to fix it.
And those five people would be the "sample". Here the people above 1200 are not the sample, they are the entire population of people above 1200.
exactly. This data is like giving the pill to everyone in the world, not just those 5 people...and the graph shows the result.
*edit/suggestion: OP, at the top of the post, state in all caps (I'll do this for you): THE DATA COLLECTED CONTAINS ALL DIAMOND PLAYERS WITH >600 POINTS.
|
On September 02 2010 08:38 GoBackToGo wrote:Show nested quote +On September 02 2010 08:32 Wr3k wrote:On September 02 2010 07:55 Liquid`Tyler wrote:
As far as I know, sc2ranks is pretty damn comprehensive of Diamond, especially high Diamond (where it seems some people have a problem with the "sample size"). Yeah, the number of people in the 1500+ group is small but that doesn't mean that there is a problem with the sample size. These numbers aren't extrapolated from a small population of the 1500+ Diamond group. These numbers directly represent that group. You would think people would realize this. It makes me cringe when they cry about sample size when ITS THE ENTIRE POPULATION. so lets say im inventing a new pill to cure something, and i test it on a handful of ppl, lets say 5.. then those 5 ppl i gave my pill to make up the whole "population" the pill was ever tested on. so lets say the test showed no adverse reaction for any of the 5 ppl. given that information, would u like to go ahead and try my new pill? btw: t is slightly imbalanced. they are going to fix it.
I don't know if you actually read all the text on the graph before you typed this post, but you should have.
The numbers represent ALL >600 Diamond players across ALL realms. There are no other people to include in the graph. There's no more diamond players because the graph already accounts for ALL of them >600.
|
On September 02 2010 08:27 arthur wrote:
if you include 1600+
then Protoss is the highest, with 3 protoss, 2 terran and 1 zerg being in the 1600's.
also note, the top 5 go zerg protoss terran protoss protoss terran, so i really dont think theres any imbalance other than how you can go semifast reaper and only be 1-2 scv's behind your opponent and have an awesome, often endgaming tool to harrass with. which is getting nerfed soon. so quit whining everyone terran is not OP.
exactly, don't know why no one noted that. the OP is withholding info to make it look like a trend.
|
On September 02 2010 08:21 MamiyaOtaru wrote:
I don't know how anyone can read this as anything but evidence that Terran is OP. Automatic match making keeps everyone at around %50 win ratio everywhere but the very top.
Mhm, i read that thread about why everyone HAS to be around 50% win ratio because of match making, but no one could explain me, why all races also have nearly the same average points. Maybe you can, i would like to hear your reasoning.
Also i want to point out, that the op coincidentally forgot to include the 1600+ column:
Protoss: 40%
Terran: 40%
Zerg: 20%
Oh, and there are 52 players between 1400 and 1600 (someone asked). And everyone talking about chi square and stuff is wrong, you cant do this kind of calculations for pure statistics like this, because we only have one measurement. You need more than one. What we could do is compare the different regions and calculate a chi square, standard derivation or whatever you want, but then you would have an answer to a different question (are the regions equal or not).
What we would need are those statistics for different point of times. Then we were at least able to say if terran has a significant dominance over time.
|
On September 02 2010 08:44 TitleRug wrote:Show nested quote +On September 02 2010 08:27 arthur wrote:
if you include 1600+
then Protoss is the highest, with 3 protoss, 2 terran and 1 zerg being in the 1600's.
also note, the top 5 go zerg protoss terran protoss protoss terran, so i really dont think theres any imbalance other than how you can go semifast reaper and only be 1-2 scv's behind your opponent and have an awesome, often endgaming tool to harrass with. which is getting nerfed soon. so quit whining everyone terran is not OP. exactly, don't know why no one noted that. the OP is withholding info to make it look like a trend.
Because if the group only 5 people in it, if you run it through some basic test for statistical significance it would not show anything due to extremely high variance.
|
On September 02 2010 08:44 TitleRug wrote:Show nested quote +On September 02 2010 08:27 arthur wrote:
if you include 1600+
then Protoss is the highest, with 3 protoss, 2 terran and 1 zerg being in the 1600's.
also note, the top 5 go zerg protoss terran protoss protoss terran, so i really dont think theres any imbalance other than how you can go semifast reaper and only be 1-2 scv's behind your opponent and have an awesome, often endgaming tool to harrass with. which is getting nerfed soon. so quit whining everyone terran is not OP. exactly, don't know why no one noted that. the OP is withholding info to make it look like a trend.
Because it is a trend. All those proportions, even the ones that are not 1300+ are statistically significant*. If you divided 1600+ out from 1500-1599, then that last group would not be conclusive either way. The others still would be.
*I have not done the math, but someone did in a previous post and got significance to a high degree of confidence for one aspect of the data.
|
1: If the discrepancy we see at the high end was just a factor of too small a data set then we would expect the graph to "jump around" a bit between each step. Instead we are seeing what appears to be a gradual shift which is highly uncharacteristic of a "too small" data set.
2: I'm absolutely not in a position to determine if one race is stronger than the others but if this was the case then this type of graph is exactly what you would expect to see. Let's say one race was buffed to be grossly OP, then in general the players of this race might end up with a rank a few hundred points above where they would otherwise be. So on average the players of this race would get "bumped up" the ladder. For most of the ladder you would not see a change at all since these players are all shifted upwards and so to speak taking each others places. The only place where it would be really noticable is at the top, since these players can't really get much higher they will accumulate and dominate the top of the ladder.
Again, I'm not saying there is an imbalance, that's for people far more knowledgeable than me to say, just that statistically this kind of graph is exactly what you would expect.
|
wow so much fail in the people posting about sample size holy cow
|
On September 02 2010 08:46 Grummler wrote:Show nested quote +On September 02 2010 08:21 MamiyaOtaru wrote:
I don't know how anyone can read this as anything but evidence that Terran is OP. Automatic match making keeps everyone at around %50 win ratio everywhere but the very top. Mhm, i read that thread about why everyone HAS to be around 50% win ratio because of match making, but no one could explain me, why all races also have nearly the same average points. Maybe you can, i would like to hear your reasoning. Also i want to point out, that the op coincidentally forgot to include the 1600+ column: Protoss: 40% Terran: 40% Zerg: 20% Oh, and there are 52 players between 1400 and 1600 (someone asked). And everyone talking about chi square and stuff is wrong, you cant do this kind of calculations for pure statistics like this, because we only have one measurement. You need more than one. What we could do is compare the different regions and calculate a chi square, standard derivation or whatever you want, but then you would have an answer to a different question (are the regions equal or not). What we would need are those statistics for different point of times. Then we were at least able to say if terran has a significant dominance over time.
Actually we have 52 measurements for players between 1400 and 1600. It is sound to do a chi-squared test of goodness of fit to assess the fit of the observed distribution to a theoretical distribution.
|
On September 02 2010 08:44 TitleRug wrote:Show nested quote +On September 02 2010 08:27 arthur wrote:
if you include 1600+
then Protoss is the highest, with 3 protoss, 2 terran and 1 zerg being in the 1600's.
also note, the top 5 go zerg protoss terran protoss protoss terran, so i really dont think theres any imbalance other than how you can go semifast reaper and only be 1-2 scv's behind your opponent and have an awesome, often endgaming tool to harrass with. which is getting nerfed soon. so quit whining everyone terran is not OP. exactly, don't know why no one noted that. the OP is withholding info to make it look like a trend.
So you're saying that if we included the 1600+ point region, that would nullify any trend that could be inferred from the graph? Sigh, please reevaluate your statement.
Here's a hint: a trend is the general movement or tendency to move in one direction. The trend wouldn't change if we added the 1600+ region.
|
Okay, I'm looking at various groupings of the Top Players in the world, and these are the stats (excluding random):
Top 5000 Players in the World - 33.57% (1687) Terran, 25.19% (1266) Zerg, 36.22% (1820) Protoss
Top 2500 - 36.38 (911) T, 23.80 (596) Z, 36.50 (914) P
Top 1000 - 39.80 (398) T, 21.30 (213) Z, 36.90 (369) P
Top 500 - 43.63 (219) T, 20.72 (104) Z, 34.06 (171) P
Top 250 - 51.20 (128) T, 18.80 (47) Z, 28.00 (70) P
Top 100 - 50.50 (51) T, 16.83 (17) Z, 29.70 (30) P
This is essentially the same information the graph shows (supports the same idea) just presented in a manner that the lesser-minded might understand.
Regardless of what the balance issue is at the moment, it is very obvious that Terran is dominating at the highest level, and only an idiot would try to dispute that.
Can we attribute these statistics to overall popularity of the race? Extremely doubtful, considering out of the ~65,000 players in Diamond (Global) Protoss is the most popular with 35%, while Terran is second at 31%. If there were a direct correlation between popularity and overall rank, Protoss should be at the very least tied with Terran at the higher levels, if not leading.
|
On September 02 2010 08:41 Flipluck wrote:
Hmmm really makes you think. I wonder whats going to happen after patch 1.1, will the reaper and tank nerf really make a difference?
Hopefully more zerg throughout the diamond league by the end of September.
Only in TvT 
I really dont think that the patch will make a huge difference. The bio ball will be unchanged, and a 5 sec later zealot will ruin the ZvP matchup, and maby even TvP. How will they defend a zergling rush on two player maps that are really hard to defend now. It will be standard to go 10 gate i suppose..
|
What I was trying to say is technically if the graph continued it would be at 1600-1700 40% and than at 2000-2100 0% while zerg is 100%
|
On September 02 2010 08:53 TitleRug wrote:
What I was trying to say is technically if the graph continued it would be at 1600-1700 40% and than at 2000-2100 0% while zerg is 100%
http://en.wikipedia.org/wiki/Outlier
|
On September 02 2010 08:52 Pekkz wrote:Show nested quote +On September 02 2010 08:41 Flipluck wrote:
Hmmm really makes you think. I wonder whats going to happen after patch 1.1, will the reaper and tank nerf really make a difference?
Hopefully more zerg throughout the diamond league by the end of September. Only in TvT I really dont think that the patch will make a huge difference. The bio ball will be unchanged, and a 5 sec later zealot will ruin the ZvP matchup, and maby even TvP. How will they defend a zergling rush on two player maps that are really hard to defend now. It will be standard to go 10 gate i suppose..
I'm kinda annoyed by this "omg I can't stop 6 pool anymore b/c of the 5 seconds" (sorry if you were refering to that, but there has been so many people QQ'ing about it). If you see a 6 pool, you WILL chrono your gate....so omg, your zealot will come out a second later...Really? (I want someone to elaborate more on this...)
|
On September 02 2010 08:53 TitleRug wrote:
What I was trying to say is technically if the graph continued it would be at 1600-1700 40% and than at 2000-2100 0% while zerg is 100%
You're trying to infer a trend based on one sample? Sigh.
|
On September 02 2010 08:48 StarDrive wrote:Show nested quote +On September 02 2010 08:46 Grummler wrote:On September 02 2010 08:21 MamiyaOtaru wrote:
I don't know how anyone can read this as anything but evidence that Terran is OP. Automatic match making keeps everyone at around %50 win ratio everywhere but the very top. Mhm, i read that thread about why everyone HAS to be around 50% win ratio because of match making, but no one could explain me, why all races also have nearly the same average points. Maybe you can, i would like to hear your reasoning. Also i want to point out, that the op coincidentally forgot to include the 1600+ column: Protoss: 40% Terran: 40% Zerg: 20% Oh, and there are 52 players between 1400 and 1600 (someone asked). And everyone talking about chi square and stuff is wrong, you cant do this kind of calculations for pure statistics like this, because we only have one measurement. You need more than one. What we could do is compare the different regions and calculate a chi square, standard derivation or whatever you want, but then you would have an answer to a different question (are the regions equal or not). What we would need are those statistics for different point of times. Then we were at least able to say if terran has a significant dominance over time. Actually we have 52 measurements for players between 1400 and 1600. It is sound to do a chi-squared test of goodness of fit to assess the fit of the observed distribution to a theoretical distribution.
What did you measure then? Because if you measure the race distribution, its only one measurement. Sure its 52 players, but only one race distribution. Sure you could count them several times, but then you would have answers to different questions, like i already said.
I mean, srsly ( i made those numbers up):
Terran: 55%
Protoss: 25%
Zerg: 20%
whats the standard deriviation? chi sqare? If anyone can tell me, i would be pretty surprised (cause you cant)
Just in case: if someone calculates the average and its variance, i will ignore him.
It is sound to do a chi-squared test of goodness of fit to assess the fit of the observed distribution to a theoretical distribution.
Yeah. Someone would think you know what you are talking about, if he never heard about statistics. But i REALLY would like to see how you compare a fit of the observed starcraft 2 race distribution to a theoretical starcraft 2 race distribution. Never heared about one.
|
On September 02 2010 08:59 seiferoth10 wrote:Show nested quote +On September 02 2010 08:53 TitleRug wrote:
What I was trying to say is technically if the graph continued it would be at 1600-1700 40% and than at 2000-2100 0% while zerg is 100% You're trying to infer a trend based on one sample? Sigh.
edit: nvm, I read something incorrectly.
On September 02 2010 09:11 Antisocialmunky wrote:
I would be more interesting in win/loss statisics in each match for players in diamond. All this shows are general trends in race popularity at various points.
|
Alternative interpretation of the data: zerg is really boring to play. For the high level, Terran has a lot more options making it more fun for skilled players, thus the better someone is, the more likely they are to play Terran, and less likely to play zerg. Similarly, mildly skilled players will find protoss the funnest because they are fairly straightforward, easier to control (read: larger units = easier to click on), but still have a reasonable amount of variety.
Summary: Currently more skilled players are less likely to zerg as their race than any other race.
|
On September 02 2010 08:58 EliteAzn wrote:Show nested quote +On September 02 2010 08:52 Pekkz wrote:On September 02 2010 08:41 Flipluck wrote:
Hmmm really makes you think. I wonder whats going to happen after patch 1.1, will the reaper and tank nerf really make a difference?
Hopefully more zerg throughout the diamond league by the end of September. Only in TvT I really dont think that the patch will make a huge difference. The bio ball will be unchanged, and a 5 sec later zealot will ruin the ZvP matchup, and maby even TvP. How will they defend a zergling rush on two player maps that are really hard to defend now. It will be standard to go 10 gate i suppose.. I'm kinda annoyed by this "omg I can't stop 6 pool anymore b/c of the 5 seconds" (sorry if you were refering to that, but there has been so many people QQ'ing about it). If you see a 6 pool, you WILL chrono your gate....so omg, your zealot will come out a second later...Really? (I want someone to elaborate more on this...)
7-8 pool allready can be a bitch to deal on two player maps. Watch madfrog vs whitera game 3 from esl.
|
|
|
Grummler, you have no idea what you are talking about.
Also I like how you ninja edited your previous post. At first, you actually had measurements, and now you have only percentages, which are meaningless. Yes, good job. You still have no idea what you're talking about.
This data is exactly analagous to the following experiment:
I have a coin, I want to see if it is fair or not. I flip the coin 50 times and record the results. I get:
Heads : 35
Tails : 15
Can I make any inference about the fairness of the coin?
You say I can't. But obviously I can. I can test the hypothesis that the coin is fair. You are full of it. The exacty same reasoning tools can be applied to the data about racial distribution at the top level. Can I ask and get a statistical answer to the question "Are the racial disparities solely due to random chance?", yes.
|
On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200.
The problem with this analysis is an incorrect null hypothesis. Why should we be assuming that 1/3 of the people prefer terran? Yes that is the random chance amount. However there is no reason to assume that. Terrans do have a reputation right now as being the easier race to learn and win with so it also is logical to skew their percentage higher.
All this graph shows is that there are more terrans in the higher diamond point range. This could be due to imbalance of course. It could also be that terran is just easier to learn (which btw is not imbalance since it just means that a Z/P player has to play for longer and train more than a terran to hit a skill cap). Could be that more people just like terran.
It requires more research.
On September 02 2010 09:06 blacktoss wrote:
Grummler, you have no idea what you are talking about.
Also I like how you ninja edited your previous post. At first, you actually had measurements, and now you have only percentages, which are meaningless. Yes, good job. You still have no idea what you're talking about.
This data is exactly analagous to the following experiment:
I have a coin, I want to see if it is fair or not. I flip the coin 50 times and record the results. I get:
Heads : 35
Tails : 15
Can I make any inference about the fairness of the coin?
You say I can't. But obviously I can. I can test the hypothesis that the coin is fair. You are full of it. The exacty same reasoning tools can be applied to the data about racial distribution at the top level. Can I ask and get a statistical answer to the question "Are the racial disparities solely due to random chance?", yes.
^^ to this post. You can test it against random chance. But that does not allow you to make the conclusion that the disparity is due to racial imbalance. Obviously this is a possibility, and likely, but it doesn't prove anything.
|
1. If you win a lot, you are removed from the lower rankings and promoted to the higher rankings.
2. To be in the highest rankings, you must consistently win a lot.
3. In the highest ranking tiers, there are more terrans than all other three races combined
4. The percentages also differ greatly from all other ranking tiers and from the overall average in favor of terran.
So those tiered percentages not only represent a current state, but a development, a history.
There are only 2 explanations for the discrepancies:
a) the best players like playing terran a lot more than all other players
b) Playing terran gives you consistently more wins compared to the other races, especially at the highest level.
And it seems to be consensus, that everyone below 1000 points is more or less a casual player.
And it also should be noted that if the lower ranks have balanced winning percentages per race, it doesn't mean that much. Because those that benefit from imbalances quickly disappear from the lower ranks. The process of promotion and demotion enforces balanced outcomes in the lower ranks. The discrepancies are passed upwards, until there is no upwards anymore where it can be passed to, and there they start to accumulate.
|
On September 02 2010 08:51 ryanAnger wrote:
Okay, I'm looking at various groupings of the Top Players in the world, and these are the stats (excluding random):
Top 5000 Players in the World - 33.57% (1687) Terran, 25.19% (1266) Zerg, 36.22% (1820) Protoss
Top 2500 - 36.38 (911) T, 23.80 (596) Z, 36.50 (914) P
Top 1000 - 39.80 (398) T, 21.30 (213) Z, 36.90 (369) P
Top 500 - 43.63 (219) T, 20.72 (104) Z, 34.06 (171) P
Top 250 - 51.20 (128) T, 18.80 (47) Z, 28.00 (70) P
Top 100 - 50.50 (51) T, 16.83 (17) Z, 29.70 (30) P
This is essentially the same information the graph shows (supports the same idea) just presented in a manner that the lesser-minded might understand.
Regardless of what the balance issue is at the moment, it is very obvious that Terran is dominating at the highest level, and only an idiot would try to dispute that.
Can we attribute these statistics to overall popularity of the race? Extremely doubtful, considering out of the ~65,000 players in Diamond (Global) Protoss is the most popular with 35%, while Terran is second at 31%. If there were a direct correlation between popularity and overall rank, Protoss should be at the very least tied with Terran at the higher levels, if not leading.
Can you add the number of randoms to that please? Then I can run a statistical test on it. I want the actual number observed, or where would I find it?
|
On September 02 2010 08:13 Cade wrote:Show nested quote +On September 02 2010 08:01 Backpack wrote:
Popularity does not equal strength.
The fact that people have this *idea* that terran might be OP and that zerg might suck will make them play terran more. You see terran all the time at high levels simply because there are more terran than zerg. The more people cry "terran OP," the more people will switch over to it.
The top players are there because they are the best players. Not because of their race. The reason there are so many terrans at the top is because there are so many terrans overall. Your post is completely misinformed, and while this logic might hold true for the lower ranges (where terran is actually NOT overly popular, protoss is the most popular. Your logic is actually very flawed when concerning top level players
Did you and everyone else who quoted me saying similar things miss the part where i said people switch?
When Dimaga says "well, guess i'll play terran" it causes tons of other players to think that they should too. Top level players are the most likely to think that they need to switch races so they can move up to the next level (tournaments and what not.)
I know P is the most popular race on the whole diamond ladder but i'm talking about even higher level than that. At MLG there were 3 zergs out of 64 players. Without even taking skill into account the chances of seeing a zerg in the finals are terrible. Whenever a player switches from Z to T, you have one less zerg player at the top and one more terran.
Doesn't anyone find it odd that the zerg race keeps getting progressively worse overtime when there haven't even been any balance changes? When we look at players like IdrA, we see that if you stick to Zerg and keep on practicing, you will still be able to compete with the best. You don't see IdrA switching to T.
Which leads back into the mentality people have that switching from Z to T will make them better and that confidence will quite possibly make them perform better. All the Z who are left just whine and complain and don't put their heart into it anymore while the newly recruited T players are excited to go beat noobs with the "OP" race.
***I'm not saying this is what's going on, since nobody can say for sure, but i just want to show people that balance isn't as bad as they think it is. This could be (and i think it is) just a social problem.
|
On September 02 2010 08:52 Pekkz wrote:Show nested quote +On September 02 2010 08:41 Flipluck wrote:
Hmmm really makes you think. I wonder whats going to happen after patch 1.1, will the reaper and tank nerf really make a difference?
Hopefully more zerg throughout the diamond league by the end of September. Only in TvT I really dont think that the patch will make a huge difference. The bio ball will be unchanged, and a 5 sec later zealot will ruin the ZvP matchup, and maby even TvP. How will they defend a zergling rush on two player maps that are really hard to defend now. It will be standard to go 10 gate i suppose..
agreed. Their PF with turret walls and their Terran Ball are being unchanged so the ZvT matchup isnt changing at all imo.
|
On September 02 2010 08:59 Grummler wrote:Show nested quote +On September 02 2010 08:48 StarDrive wrote:On September 02 2010 08:46 Grummler wrote:On September 02 2010 08:21 MamiyaOtaru wrote:
I don't know how anyone can read this as anything but evidence that Terran is OP. Automatic match making keeps everyone at around %50 win ratio everywhere but the very top. Mhm, i read that thread about why everyone HAS to be around 50% win ratio because of match making, but no one could explain me, why all races also have nearly the same average points. Maybe you can, i would like to hear your reasoning. Also i want to point out, that the op coincidentally forgot to include the 1600+ column: Protoss: 40% Terran: 40% Zerg: 20% Oh, and there are 52 players between 1400 and 1600 (someone asked). And everyone talking about chi square and stuff is wrong, you cant do this kind of calculations for pure statistics like this, because we only have one measurement. You need more than one. What we could do is compare the different regions and calculate a chi square, standard derivation or whatever you want, but then you would have an answer to a different question (are the regions equal or not). What we would need are those statistics for different point of times. Then we were at least able to say if terran has a significant dominance over time. Actually we have 52 measurements for players between 1400 and 1600. It is sound to do a chi-squared test of goodness of fit to assess the fit of the observed distribution to a theoretical distribution. What did you measure then? Because if you measure the race distribution, its only one measurement. Sure its 52 players, but only one race distribution. Sure you could count them several times, but then you would have answers to different questions, like i already said. I mean, srsly ( i made those numbers up): Terran: 55% Protoss: 25% Zerg: 20% whats the standard deriviation? chi sqare? If anyone can tell me, i would be pretty surprised (cause you cant) Just in case: if someone calculates the average and its variance, i will ignore him.
I'm not really sure how to address all the issues here but the misunderstandings you're having are fairly subtle and many people who don't have a super strong statistics background have them so this isn't a simple case of someone completely misunderstanding statistics.
The first thing is in your interpretation of one sample, then if we measured the height of all males in the US to compute the average height, by your argument, that would also be one sample since there is only one US. To see more samples we would have to measure another instance of the universe. You can think of this loosely as a parallel universe, although this is not rigorous by any means.
You ask for the standard deviation which is a fair question. That is an important measure of a distribution and is important to assess the normal approximation of the binomial as I did in an earlier post. However, you did not specify which distribution you want the standard deviation of.
I will not go through all the theory and methodology of performing a chi-square goodness of fit test on a distribution here. But it suffices to say that it is sound and accepted practice. You can find more information here http://en.wikipedia.org/wiki/Pearson's_chi-square_test
I think if you gave your object of there being only one trend a bit more thought, you'd see that it doesn't really make sense to require multiple samples of distributions (not samples from a distribution) to perform this test. In that case, what would you be doing your fit on?
|
I would be more interesting in win/loss statisics in each match for players in diamond. All this shows are general trends in race popularity at various points.
|
On September 02 2010 08:58 EliteAzn wrote:Show nested quote +On September 02 2010 08:52 Pekkz wrote:On September 02 2010 08:41 Flipluck wrote:
Hmmm really makes you think. I wonder whats going to happen after patch 1.1, will the reaper and tank nerf really make a difference?
Hopefully more zerg throughout the diamond league by the end of September. Only in TvT I really dont think that the patch will make a huge difference. The bio ball will be unchanged, and a 5 sec later zealot will ruin the ZvP matchup, and maby even TvP. How will they defend a zergling rush on two player maps that are really hard to defend now. It will be standard to go 10 gate i suppose.. I'm kinda annoyed by this "omg I can't stop 6 pool anymore b/c of the 5 seconds" (sorry if you were refering to that, but there has been so many people QQ'ing about it). If you see a 6 pool, you WILL chrono your gate....so omg, your zealot will come out a second later...Really? (I want someone to elaborate more on this...)
I'm not really a protoss player so I'm not sure exactly on the issue so correct me if I'm wrong, but I think it is that p has to cut probes to defend against early pool because of the zealot nerf, so they have to cut probes all the time because of the threat of an early pool but if the opponent went standard they would be economically behind.
|
On September 02 2010 09:11 Antisocialmunky wrote:
I would be more interesting in win/loss statisics in each match for players in diamond. All this shows are general trends in race popularity at various points.
Players in diamond will by definition have similar win loss statistics. But the trend is clear: strong players prefer Terran.
|
On September 02 2010 09:06 blacktoss wrote:
Grummler, you have no idea what you are talking about.
Also I like how you ninja edited your previous post. At first, you actually had measurements, and now you have only percentages, which are meaningless. Yes, good job. You still have no idea what you're talking about.
This data is exactly analagous to the following experiment:
I have a coin, I want to see if it is fair or not. I flip the coin 50 times and record the results. I get:
Heads : 35
Tails : 15
Can I make any inference about the fairness of the coin?
You say I can't. But obviously I can. I can test the hypothesis that the coin is fair. You are full of it. The exacty same reasoning tools can be applied to the data about racial distribution at the top level. Can I ask and get a statistical answer to the question "Are the racial disparities solely due to random chance?", yes.
No, you cant. Cause you totaly ignore the amount of ppl playing that race. And i "ninja edited" my post, cause the op gives us no absolut numbers like i did in my example. Just because you understood that high school math example doesnt prove anything.
|
@EliteAzn the raw data can be found in the link I gave in my first post. It is clear how I did the numbers (except for one little thing, the 1500 is in fact 1500+ because there are only 9 above 1600). Anyhow here is the data.
|
Alternative explanation: Racial differential in skill scaling:
Terran has the most mechanics that have higher skill caps, i.e., to master (get full potential) out of the Terran, it requires more skill than the other races. (Followed by toss, then zerg.)
This means that since 99.99% of people don't have the skill to reach that level, Blizzard balancing for most people leaves mostly Terrans at the top.
Solution: Make the other races more interesting by adding skill-differentiating abilities - don't just nerf and buff units. E.g., make roaches interesting again by increasing their regen and lowering their hp; Add an early game spell to zerg like toss has with sentry; make sentries do 8 dmg again.
|
On September 02 2010 09:12 Grummler wrote:Show nested quote +On September 02 2010 09:06 blacktoss wrote:
Grummler, you have no idea what you are talking about.
Also I like how you ninja edited your previous post. At first, you actually had measurements, and now you have only percentages, which are meaningless. Yes, good job. You still have no idea what you're talking about.
This data is exactly analagous to the following experiment:
I have a coin, I want to see if it is fair or not. I flip the coin 50 times and record the results. I get:
Heads : 35
Tails : 15
Can I make any inference about the fairness of the coin?
You say I can't. But obviously I can. I can test the hypothesis that the coin is fair. You are full of it. The exacty same reasoning tools can be applied to the data about racial distribution at the top level. Can I ask and get a statistical answer to the question "Are the racial disparities solely due to random chance?", yes. No, you cant. Cause you totaly ignore the amount of ppl playing that race. And i "ninja edited" my post, cause the op gives us no absolut numbers like i did in my example. Just because you understood that high school math example doesnt prove anything.
Grummler, we can deduce the actual numbers by looking at the site ourselves, so we can perform statistical tests. I do agree that it requires more than just percentages, but it is possible as I have shown in a previous post. No need to attack other posters. Just respond to their arguments logically.
|
On September 02 2010 09:12 Grummler wrote:Show nested quote +On September 02 2010 09:06 blacktoss wrote:
Grummler, you have no idea what you are talking about.
Also I like how you ninja edited your previous post. At first, you actually had measurements, and now you have only percentages, which are meaningless. Yes, good job. You still have no idea what you're talking about.
This data is exactly analagous to the following experiment:
I have a coin, I want to see if it is fair or not. I flip the coin 50 times and record the results. I get:
Heads : 35
Tails : 15
Can I make any inference about the fairness of the coin?
You say I can't. But obviously I can. I can test the hypothesis that the coin is fair. You are full of it. The exacty same reasoning tools can be applied to the data about racial distribution at the top level. Can I ask and get a statistical answer to the question "Are the racial disparities solely due to random chance?", yes. No, you cant. Cause you totaly ignore the amount of ppl playing that race. And i "ninja edited" my post, cause the op gives us no absolut numbers like i did in my example. Just because you understood that high school math example doesnt prove anything.
http://sc2ranks.com/stats/race/all/1
Right on the original post. Yes, yes they did.
And, once again. Yes you can. There are two possibilities here. Either you did not click on the most important link in this thread (the one with the actual data on it), or you do not understand statistics at all. Or both. Which is it?
|
On September 02 2010 09:07 Backpack wrote:
***I'm not saying this is what's going on, since nobody can say for sure, but i just want to show people that balance isn't as bad as they think it is. This could be (and i think it is) just a social problem.
I'm not gonna quote the whole thing as I'm afraid it might be blown out of context, but this tidbit here is quite interesting.
It could be just a social aspect of the Sc2 community right now for players to switch from Z to T. I'm not saying that I'm a pro or anything but to me the balances are as severe as people are making them out to be, I'm having a ton of trouble with Reapers yeah but beyond that I feel I can compete with the Terran, albeit even if it feels like I'm putting more effort into my part.
What am I trying to get at? Nothing really, but it might be interesting to see wether or not players switch out of T soon after the patch JUST because of the fact that they are getting nerfed, not the fact that the nerf is effecting the race so heavily.
|
@TitleRug I did not include the 1600 intervall because I thought it would be to small since it would contain five entries. I thought about calling the last intervall 1500+ but thought it would be to much for the a post in the internet.
|
On September 02 2010 09:18 ReplayArk wrote:
@TitleRug I did not include the 1600 intervall because I thought it would be to small since it would contain five entries. I thought about calling the last intervall 1500+ but thought it would be to much for the a post in the internet.
Fair enough
|
LOLOL hahaha
Everybody likes the terran 
|
Only a basic understanding of statistics and logic is required to understand this information, and yet so many people fail. (Pretty much everyone but the "pill" guy)
1. Popularity is not identical to "overpoweredness". This is a BIG assumption. More top players prefer terran, that's true, but knowing why they do so is infinitely more informative than just the population data. If they all say "I play terran because I do better" (and we have reliable confidence they have no reason to deceive us) then sure, get up on your soapbox. But it won't be because of this graph, it will be because of the top players' reasoning. For all you know, they might say "I like challenge" (to play devil's advocate).
2. To settle the sample size debate once and for all: The top players cannot be their own population if you are trying to make a generalization about the GAME. You can make a generalization about top players or top end play, if you like. That's what samples do, they generalize. So it completely depends on what exactly you are trying to generalize. It will be too small for some generalizations, and plenty large enough for others. It's not one or the other, it depends what you are trying to argue.
|
On September 02 2010 09:21 ToxNub wrote:
Only a basic understanding of statistics and logic is required to understand this information, and yet so many people fail. (Pretty much everyone but the "pill" guy)
1. Popularity is not identical to "overpoweredness". This is a BIG assumption. More top players prefer terran, that's true, but knowing why they do so is infinitely more informative than just the population data. If they all say "I play terran because I do better" (and we have reliable confidence they have no reason to deceive us) then sure, get up on your soapbox. But it won't be because of this graph, it will be because of the top players' reasoning. For all you know, they might say "I like challenge" (to play devil's advocate).
2. To settle the sample size debate once and for all: The top players cannot be their own population if you are trying to make a generalization about the GAME. You can make a generalization about top players or top end play, if you like. That's what samples do, they generalize. So it completely depends on what exactly you are trying to generalize. It will be too small for some generalizations, and plenty large enough for others. It's not one or the other, it depends what you are trying to argue.
Yes it could be that they play Terran since they like a challenge and they are so much more talented than everyone else that they can dominate even with the additional challenge. However, this explanation seems unlikely.
|
On September 02 2010 09:21 ToxNub wrote:
Only a basic understanding of statistics and logic is required to understand this information, and yet so many people fail. (Pretty much everyone but the "pill" guy)
2. To settle the sample size debate once and for all: The top players cannot be their own population if you are trying to make a generalization about the GAME. You can make a generalization about top players or top end play, if you like. That's what samples do, they generalize. So it completely depends on what exactly you are trying to generalize. It will be too small for some generalizations, and plenty large enough for others. It's not one or the other, it depends what you are trying to argue.
I'm only going to respond to #2.
Um, the GAME [sic] is not a population. The GAME is what it is. The players of the GAME form a population. The data discussed in this thread is not talking about the population of all players. If it were, then it would be useless, because it is not a random sample. It is all diamond players. If you want to make inferences about a population from a subset of that population, you want to use random sampling.
But that is not the reasoning used here. The first line of reasoning is to ask whether there is any racial bias in the distribution of players in higher level diamond league that is not due to random chance. The answer is, without a shadow of reasonable doubt, yes. There is such a bias. The next question is "What causes this bias?"
And in this case, there can be many claims trying to explain the bias. Why is racial imbalance a good one? The claim is not that one is trying to generalize from a 'sample' (diamond population) to a 'population' (the GAME), it is one of measuring one variable with another.
The variable we have to work with is racial distribution in upper diamond league. The variable we are interested in is racial balance. When you try to infer one variable from another, the one you measure is called a proxy. This is not the same as an inference about a population from a sample. Please do not conflate the two.
|
On September 02 2010 09:24 StarDrive wrote:Show nested quote +On September 02 2010 09:21 ToxNub wrote:
Only a basic understanding of statistics and logic is required to understand this information, and yet so many people fail. (Pretty much everyone but the "pill" guy)
1. Popularity is not identical to "overpoweredness". This is a BIG assumption. More top players prefer terran, that's true, but knowing why they do so is infinitely more informative than just the population data. If they all say "I play terran because I do better" (and we have reliable confidence they have no reason to deceive us) then sure, get up on your soapbox. But it won't be because of this graph, it will be because of the top players' reasoning. For all you know, they might say "I like challenge" (to play devil's advocate).
2. To settle the sample size debate once and for all: The top players cannot be their own population if you are trying to make a generalization about the GAME. You can make a generalization about top players or top end play, if you like. That's what samples do, they generalize. So it completely depends on what exactly you are trying to generalize. It will be too small for some generalizations, and plenty large enough for others. It's not one or the other, it depends what you are trying to argue.
Yes it could be that they play Terran since they like a challenge and they are so much more talented than everyone else that they can dominate even with the additional challenge. However, this explanation seems unlikely.
Sure, but this data doesn't say that. That's your opinion. That's all i'm saying.
|
This data is not saying that terrans are winning more games than zerg or protoss. This data says there are just more terran players in the highest diamond leagues. Also its not that drastic of a difference then it looks like. Terran doubled while Protoss and Zerg dropped half.
This can have many reasons:
1. People switching from P/Z to T because they think they can have an advantage because terran is said to be op
2. More super high level players play T instead of P/Z because terran is said to be the most creative race
3. At the top level are mostly players with years experience of Starcraft Brood War who transitioned easily into high level Starcraft2. Now the mostly dominant race in Brood War for the last year was Terran because of Flash. People saw that and choosed mostly Terran (even if they weren't terran in Brood War) because they wanted to learn terran. What is there a better point of time to switch to terran than the start of a new game.
|
This really doesn't mean anything to BLIZZARD, just to the competitive scene. Platinum and Diamond are both Protoss dominated leagues, and the issue with this is that they don't make up a majority of the players overall, so it actually tells you that the top end is Protoss favored with Terran making up more of the Bronze leaguers etc...
800 people at the top compared to over a million on 1v1 ladder, it would seem that P is the overpowered race allowing their general population to skew towards diamond rather than T being overpowered since T players overall skew towards the low end of the ladder...
|
On September 02 2010 09:31 ToxNub wrote:Show nested quote +On September 02 2010 09:24 StarDrive wrote:On September 02 2010 09:21 ToxNub wrote:
Only a basic understanding of statistics and logic is required to understand this information, and yet so many people fail. (Pretty much everyone but the "pill" guy)
1. Popularity is not identical to "overpoweredness". This is a BIG assumption. More top players prefer terran, that's true, but knowing why they do so is infinitely more informative than just the population data. If they all say "I play terran because I do better" (and we have reliable confidence they have no reason to deceive us) then sure, get up on your soapbox. But it won't be because of this graph, it will be because of the top players' reasoning. For all you know, they might say "I like challenge" (to play devil's advocate).
2. To settle the sample size debate once and for all: The top players cannot be their own population if you are trying to make a generalization about the GAME. You can make a generalization about top players or top end play, if you like. That's what samples do, they generalize. So it completely depends on what exactly you are trying to generalize. It will be too small for some generalizations, and plenty large enough for others. It's not one or the other, it depends what you are trying to argue.
Yes it could be that they play Terran since they like a challenge and they are so much more talented than everyone else that they can dominate even with the additional challenge. However, this explanation seems unlikely. Sure, but this data doesn't say that. That's your opinion. That's all i'm saying.
Well by this argument then all the data in support of General Relativity could have been manipulated by aliens to make it look like General Relativity provides a good model of the universe when in reality Newtonian mechanics is better. The data does not disprove that hypothesis so to anyone who believes otherwise: "It's just your opinion."
|
On September 02 2010 09:14 NATO wrote:
Alternative explanation: Racial differential in skill scaling:
I think this has very little to do with skill scaling.
It's simply a result of promoting for wins and demoting for losses. So someone who wins a lot quickly disappears from the lower ranks, and in the lower ranks the win percentages necessarily are always near the perfect balance. That's the point of the whole system.
At the top there is no more promotion, so any discrepancy will accumulate there. When there is a group that consistently adds up more wins than everyone else, there will be more of that group at the top. And the most obvious group that consistently gets more victories than everyone else is terrans.
And just the fact that they are terrans is a sufficient explanation. And it's the only explanation that the data allows. Because every other explanation needs data that is not available, and frankly is a hard sell with a lot of unreasonable assumptions.
Now what exactly causes this high terran win rate, that can't be swept under the rug, that is a a different question. All you can say for now, that it is more likely/easier for terrans to win.
|
You're never going to see elevated win percentages really due to the way the matchmaking system works. What you will see is elevated ratings because the "cushion" is effectively your rating, if your winning too often your rating climbs and your matched up against people far above your true rating until you start to lose, forcing you closer to 50%. So this means since the system is designed to jack your rating up as high as required to keep you as close to 50% as possible, ratings will tell the true story of balance as opposed to win percentages. So the fact that we see so many Terran's at the highest level of rating says to me either 1) all Terran players are just better players than everyone else, or 2) it is in fact easier to win with the Terran race. I think #1 will be a more temporary possibility as other players learn, practice, and get better. However #2 is something that just won't go away, in fact it will grow throughout the player population as everyone's skill level starts to grow.
Only time will tell and I imagine Blizzard, having a lot of experience in balancing games, knows this and they are waiting to do anything drastic. So its no surprise 1.1 is a relatively small patch, if these kind of numbers continue though for say 6 months I bet they start making some pretty sweeping changes.
|
On September 02 2010 08:31 Mortecian wrote:
It means that at the highest level of play, there are more terrans than all other races combined.
If you are looking at the whole population, for instance, let's say that population is "Highest level of play", then there is no need to speak of sample size... you're looking at the whole population, there is no estimate.
Is this totally indicative that there is imbalance? No. But it supports that idea.
- If many experts are saying terran is imba, it again supports it.
- If blizzard says terran is imba and will tweak in 1.1, it again supports it.
- If experts in other races do not refute that terran is imba, it again supports it.
So, saying that at the top level of competitive play, terran is imbalanced is supported by:
- Data above.
- Experts opinion (Idra, Sheth, Morrow, etc. etc.).
- Designers opinion (Blizzard Patch 1.1 status update).
Whether it's true or not, I'm not really vested in or against... as I dont play starcraft 2, but these three sources all point toward the same idea and that gives the idea weight.
So many signs pointing to one thing, but people still will defend it until their deaths. its amazing.
|
On September 02 2010 09:27 blacktoss wrote:Show nested quote +On September 02 2010 09:21 ToxNub wrote:
Only a basic understanding of statistics and logic is required to understand this information, and yet so many people fail. (Pretty much everyone but the "pill" guy)
2. To settle the sample size debate once and for all: The top players cannot be their own population if you are trying to make a generalization about the GAME. You can make a generalization about top players or top end play, if you like. That's what samples do, they generalize. So it completely depends on what exactly you are trying to generalize. It will be too small for some generalizations, and plenty large enough for others. It's not one or the other, it depends what you are trying to argue.
I'm only going to respond to #2. Um, the GAME [sic] is not a population. The GAME is what it is. The players of the GAME form a population. The data discussed in this thread is not talking about the population of all players. If it were, then it would be useless, because it is not a random sample. It is all diamond players. If you want to make inferences about a population from a subset of that population, you want to use random sampling. But that is not the reasoning used here. The first line of reasoning is to ask whether there is any racial bias in the distribution of players in higher level diamond league that is not due to random chance. The answer is, without a shadow of reasonable doubt, yes. There is such a bias. The next question is "What causes this bias?" And in this case, there can be many claims trying to explain the bias. Why is racial imbalance a good one? The claim is not that one is trying to generalize from a 'sample' (diamond population) to a 'population' (the GAME), it is one of measuring one variable with another. The variable we have to work with is racial distribution in upper diamond league. The variable we are interested in is racial balance. When you try to infer one variable from another, the one you measure is called a proxy. This is not the same as an inference about a population from a sample. Please do not conflate the two.
It is an inference about a population from a sample, you simply word it in a way that obscures it. You want to say "the current players in top diamond have a racial bias". Yes, the data shows that. However, then you want to pretend (and this is a subtle difference) that it implies that any new top diamond player is likely to follow the distribution. You must, if you intend to propose that the game, which affects any player that uses it (and not just the current top diamond players), is responsible.
A simple example:
I flip a coin to determine the "preference" for heads or tails, let's say I do that for different brackets. But for each bracket I flip it a different number of times.
1300 (10 flips): 5:5
1400 (5 flips): 3:2
1500: (1 flip) 1:0
Now I could make a big graph of this, inflate the number of trials on the low end, and make a thread about coin imbalance. You would come into my thread, and you would ask "Is there any preference bias in the distribution of coin flips in higher level coin league that is not due to random chance. The answer is, without a shadow of reasonable doubt, yes."
Think about that for a second. It doesn't matter if all of the coin flips in 1500 is the entire population, you must consider a NEW sample's likelihood to follow this distribution. That is the whole point of a confidence measure.
|
On September 02 2010 09:37 StarDrive wrote:Show nested quote +On September 02 2010 09:31 ToxNub wrote:On September 02 2010 09:24 StarDrive wrote:On September 02 2010 09:21 ToxNub wrote:
Only a basic understanding of statistics and logic is required to understand this information, and yet so many people fail. (Pretty much everyone but the "pill" guy)
1. Popularity is not identical to "overpoweredness". This is a BIG assumption. More top players prefer terran, that's true, but knowing why they do so is infinitely more informative than just the population data. If they all say "I play terran because I do better" (and we have reliable confidence they have no reason to deceive us) then sure, get up on your soapbox. But it won't be because of this graph, it will be because of the top players' reasoning. For all you know, they might say "I like challenge" (to play devil's advocate).
2. To settle the sample size debate once and for all: The top players cannot be their own population if you are trying to make a generalization about the GAME. You can make a generalization about top players or top end play, if you like. That's what samples do, they generalize. So it completely depends on what exactly you are trying to generalize. It will be too small for some generalizations, and plenty large enough for others. It's not one or the other, it depends what you are trying to argue.
Yes it could be that they play Terran since they like a challenge and they are so much more talented than everyone else that they can dominate even with the additional challenge. However, this explanation seems unlikely. Sure, but this data doesn't say that. That's your opinion. That's all i'm saying. Well by this argument then all the data in support of General Relativity could have been manipulated by aliens to make it look like General Relativity provides a good model of the universe when in reality Newtonian mechanics is better. The data does not disprove that hypothesis so to anyone who believes otherwise: "It's just your opinion."
WTF are you talking about. YOU are manipulating data and providing your own indepedent opinion as a substitute for an accurate interpretation. If you have to fill in holes in the data with your opinion, then it's no longer data, it's your opinion.
|
This doesn't really prove anything. I guarantee if you did the same thing for BW you would not see an even distribution of the races (in fact, this is quite widely known). While I don't discount the notion that if a race is imbalanced, more players will play that race, the opposite (if more players play a race, then that race is imbalanced) is not always true.
|
The numbers seem even worse when you realize that 40-50% more people play Protoss than Terran. Zerg is the combination of being unpopular and weak.
|
40-50% more people play Protoss than Terran
@Cloak, I would like to see the source for this number.
|
Say what you want, this data CAN'T PROVE terran is op. It PROVES that there are more high level terran players. And there can be many many reasons for that
|
Its shocking that people are claiming that the far end of the bell curve is meaningless because it has few people in it. There are few people in it precisely because each group must be smaller than the previous, and the trends appear to hold even in the extreme case of 1500+
Whether this has to do with the perceived preference for terran or an actual one is an interesting topic that is hard to evaluate. My gut says the trend TvZ represents the array of viable openings on one side that Z just gets to suffer through until midgame.
|
this can also by protoss players are easier to play and get on high leagues and terrans harder but more consistent ....
|
Why people are choosing to ignore what Tyler says is beyond me. This sint a statistical test, this is a mapping of everyone! And as such it shows the reality of the composition of top ranked players, so don't try to sound like a smartass with your course in statistics and say something that is blatantly false and ignorant. Can you tell something from this graph? Yes! At the top of ladder there ARE less zerg! That's no statistical conclusion its the fact of the matter!
Edit: that does not mean Zerg is up.
|
Nice to read comment from people who have heard the word "sample size" before saying that it isnt large enough. He isnt using part of the population of high ranked players to show a trend for high ranked players. He is using ALL of the high ranked players. He isnt saying anything about what effect race has below diamond level, and he isnt saying what will happen to future players. He is showing what has already happened in this patch.
|
On September 02 2010 10:51 Three wrote:
Nice to read comment from people who have heard the word "sample size" before saying that it isnt large enough. He isnt using part of the population of high ranked players to show a trend for high ranked players. He is using ALL of the high ranked players. He isnt saying anything about what effect race has below diamond level.
I think by "sample size" they mean the low amount of top level players compared to the rest of the population.
|
On September 02 2010 10:52 TitleRug wrote:Show nested quote +On September 02 2010 10:51 Three wrote:
Nice to read comment from people who have heard the word "sample size" before saying that it isnt large enough. He isnt using part of the population of high ranked players to show a trend for high ranked players. He is using ALL of the high ranked players. He isnt saying anything about what effect race has below diamond level. I think by "sample size" they mean the low amount of top level players compared to the rest of the population.
Thats not what hes looking at
|
On September 02 2010 10:52 TitleRug wrote:Show nested quote +On September 02 2010 10:51 Three wrote:
Nice to read comment from people who have heard the word "sample size" before saying that it isnt large enough. He isnt using part of the population of high ranked players to show a trend for high ranked players. He is using ALL of the high ranked players. He isnt saying anything about what effect race has below diamond level. I think by "sample size" they mean the low amount of top level players compared to the rest of the population.
Right, but it's still a decent sample size. Medical studies on rats are often done with only 8-10 samples and those are generally statistically sound. Someone needs to actually crunch the uncertainties instead of letting people merely assert that the sample size is too small. If you think it's too small, do the uncertainty calculations to prove it.
|
People keep saying that that with imperfect control (1100 and below) it is easier for Protoss to win and at higher levels (1200 and above) Terran has an advantage.
Is it not equally possible that Protoss players who approach perfect control start losing to Terran players with lesser control at around the 1200 level and this just gets worse as the ratings get higher? It seems unfair to say "At lower skill levels Protoss can win more" when it seems just as likely that the figurative skill ceiling of Protoss is simply lower, causing more skilled protoss players to be left at the lower levels.
And man I feel sorry for Zerg in the TvZ match up.
|
On September 02 2010 10:54 Three wrote:Show nested quote +On September 02 2010 10:52 TitleRug wrote:On September 02 2010 10:51 Three wrote:
Nice to read comment from people who have heard the word "sample size" before saying that it isnt large enough. He isnt using part of the population of high ranked players to show a trend for high ranked players. He is using ALL of the high ranked players. He isnt saying anything about what effect race has below diamond level. I think by "sample size" they mean the low amount of top level players compared to the rest of the population. Thats not what hes looking at
maybe you're right, I lost track of what people said.
On September 02 2010 10:55 Sentient wrote:Show nested quote +On September 02 2010 10:52 TitleRug wrote:On September 02 2010 10:51 Three wrote:
Nice to read comment from people who have heard the word "sample size" before saying that it isnt large enough. He isnt using part of the population of high ranked players to show a trend for high ranked players. He is using ALL of the high ranked players. He isnt saying anything about what effect race has below diamond level. I think by "sample size" they mean the low amount of top level players compared to the rest of the population. Right, but it's still a decent sample size. Medical studies on rats are often done with only 8-10 samples and those are generally statistically sound. Someone needs to actually crunch the uncertainties instead of letting people merely assert that the sample size is too small. If you think it's too small, do the uncertainty calculations to prove it.
don't quote me, I'm just assuming that's what those people meant.
|
On September 02 2010 09:38 imbecile wrote:Show nested quote +On September 02 2010 09:14 NATO wrote:
Alternative explanation: Racial differential in skill scaling:
I think this has very little to do with skill scaling. It's simply a result of promoting for wins and demoting for losses. So someone who wins a lot quickly disappears from the lower ranks, and in the lower ranks the win percentages necessarily are always near the perfect balance. That's the point of the whole system. At the top there is no more promotion, so any discrepancy will accumulate there. When there is a group that consistently adds up more wins than everyone else, there will be more of that group at the top. And the most obvious group that consistently gets more victories than everyone else is terrans. And just the fact that they are terrans is a sufficient explanation. And it's the only explanation that the data allows. Because every other explanation needs data that is not available, and frankly is a hard sell with a lot of unreasonable assumptions. Now what exactly causes this high terran win rate, that can't be swept under the rug, that is a a different question. All you can say for now, that it is more likely/easier for terrans to win.
This should be in the OP; I don't think most people understand it
|
Add tank overkill, buff damage, nerf the marauder, add a tech building. Done.
|
On September 02 2010 09:04 NATO wrote:
Alternative interpretation of the data: zerg is really boring to play. For the high level, Terran has a lot more options making it more fun for skilled players, thus the better someone is, the more likely they are to play Terran, and less likely to play zerg. Similarly, mildly skilled players will find protoss the funnest because they are fairly straightforward, easier to control (read: larger units = easier to click on), but still have a reasonable amount of variety.
Summary: Currently more skilled players are less likely to zerg as their race than any other race.
And that would be OK? Zerg not being fun to play?
|
These results show that Random is clearly underpowered.
And that would be OK? Zerg not being fun to play?
I don't see where in his post he said that's OK. As a Random player I actually think the same thing: Zerg isn't too terribly weak in terms of balance. Patch 1.1 will resolve the current deepest imbalance of Reapers, but even that isn't too bad until you face Terrans of 1100 Diamond or higher.
However, when I draw Zerg I usually sigh because I know the same routine is coming. Speedlings into either a baneling bust if they (hopefully) leave themselves open to it, or it's time to play a macro war while the enemy P or T enjoys a significantly easier macro and micro game. Lair tech opens up a few possibilities but I feel like at that point, one wrong move by me as Zerg leads to an instant loss.
Overall playing as Zerg is one dimensional, taxing and frustrating because it's the only race I feel where I can out-macro and out-micro to a large degree and still lose to A moving, inferior players with one mistake.
|
I'm interested in why random starts off high, dips in the middle, and goes back up, in a really consistent manner too.
Anyone have a plausible explanation for this?
|
On September 02 2010 11:44 roronoe wrote:
I'm interested in why random starts off high, dips in the middle, and goes back up, in a really consistent manner too.
Anyone have a plausible explanation for this?
Players start out wanting to play all races. Than as they get better they pick one race to work on. Than at highest level they play random for surprise advantage or they try to understand and get good at all races.
|
I think the most blatantly obvious point here is that Zerg sucks.
|
On September 02 2010 08:14 blacktoss wrote:
ITT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected.
"Terran players make up a disproportionate amount of 1300+ Diamond ladder players in a way that can't be attributed to random chance." - Fact? You can't readily see this from the data. You must use something like a chi-square test to determine the confidence you can attach to this statement. I think it would be a good thing if someone did this with the raw data. Preferably someone with a more sophisticated background than one year of college statistics (moi).
"Terran players make up a disproportionate amount of 1300+ Diamond ladder players and it is because Terran is imbalanced." Fact??? This would be very hard to determine conclusively, and you can't even ascribe a confidence to this claim with the data available. However, under reasonable assumptions, it is a reasonable claim.
Other claims, like "Terran is more prevalent in higher leagues because more people play it because of the campaign" are more readily testable, but require sampling of campaign play data with respect to skill level.
I really wish people would pay more attention to this post. Because they aren't this thread is just derailing into senseless whining with little understanding of what the data do and don't tell us.
|
|
|
Chile4253 Posts
On September 02 2010 07:57 Mikilatov wrote:
Pretty eye-opening, it seems.
I'm glad that this graph pulls up an interesting point though, Terrans aren't really that overpowered except at high levels in the hands of 1000+ point diamond players.
No, that is not the correct way to read the graph.
Think of it this way: given an Auto Match-making system (AMM) that pits you against opponents whom you are about as likely to win against as you are to lose against, there should be a population spread relatively similar to the overall popularity of each race on a global level (e.g. if 25% players play Zerg, 35% play Protoss, 30% play Terran, and 10% play Random, then you should see numbers similar to those on most rank intervals).
The only place where racial imbalance should begin to make a break in this spread is in very high-level play. If, say, Terran was overpowered, then you could see Terran players who belong at Diamond 500 skill-wise ranked at Diamond 700; they are less skilled than their Protoss and Zerg counterparts at the Diamond 700 level, but the racial imbalance props them up.
However, when you have the best players of the world who are at the "skill cap" (you can define this as 1300+, 1400+, 1500+, etc), if there really is a racial imbalance that props up your rank, then that's a lot more evident. You'd expect to see something like 30% Terrans at the 1500+ level, but there are 60% instead- you'd expect about half of those are getting propped up by the racial imbalance.
TL;DR: It is easiest to see racial imbalance at the very top skill ranks, and the data suggests pretty strongly (IMO) that there is one in favor of Terran and POSSIBLY in detriment of Zerg (Zerg underpowered argument is less strong because it's only evident at the 1400+ mark, with a very small population size of 60).
|
people in this thread absolutely infuriate me, the terran imbalance bullshit is seriously to a point where i cant believe it. im not saying anything regarding balance of the game, but for fuck sake just stop fucking whining, maybe if P and Z players learned how to play the fucking game instead of spending their entire fucking day writing garbage on teamliquid they would get somewhere. lets look at a few very basic facts that require absolutely no data to be proven
#1 the first game released is from the terran campaign
#2 every fucking post in every god damn starcraft 2 forum says something about OMFG TERRAN IMBALANCED WAT 2 do OMNFG.
these 2 factors alone are going to encourage a ton more people playing terran. more people playing a race is going to lead to a greater development in strategies, leading to a more developed race
lets look at some very basic statistics for all of the illiterate 12 year old retards who dont understand how to interpret anything that doesnt tell them UR A GOOD PLAYER U JUST LOST CUS THE GAME IS IMBALANCED
number of players at a given level of play does not correlate with balance. AT ALL. lets say fayth made a post in sc2 strategy saying he would give 1000$ to every player who earned 1500 rating as zerg. boy these graphs would look alot different wouldnt they. does that mean that zerg is all of a sudden overpowered and stronger then terran? no.
the dumbest thing by far in this thread is the 1600 and 1500 rating level graph, by far. anyone whos done anything in statistics (or has half a brain) knows that a sample size of 19 (YES THERE ARE 19 PLAYERS ABOVE 1500) is not an adequate sample size to judge something as complex as game balance. IM A MORON, WHY IS THIS PETZERGLING? each 1 person on the graph represents over 5% of the total population. lets say a T player with 1501 rating plays a Z player with 1499 rating and the Z player wins (LOL CANT HAPPEN TERRAN OP) this brings an adjustment to our graph of over 10% relative data. considering people at this level play upwards of 10 games a day. proof that this sample size is retarded? here i just re calculated the numbers courtesy of sc2ranks.com and put them on a graph
![[image loading]](http://img830.imageshack.us/img830/9837/lolimbalanve.jpg)
OMFG PROTOSS UPSWING NOW PROTOSS IS IMBALANCED.
no the race of players in the top 1500 fluctuates this much on a daily(hourly?) basis
k im done see you idiots in disneyland
User was temp banned for this post.
|
Something I'd like to see is the compared win/loss rations per race for random players. Random players should have roughly equivalent experience and skill with each race, and if there is a trend towards random players doing better or worse overall with a race, that would indicate an OP/UP race imo. What made me think of this was taking a look at Danrok's stats page (http://www.danrok.com/stats/) where his win rates for T/P/Z are 56/57/50. I don't think it's possible to pull that info out of sc2ranks.com... maybe that's only info that Blizzard will know.
|
On September 02 2010 13:14 petzergling wrote:
people in this thread absolutely infuriate me, the terran imbalance bullshit is seriously to a point where i cant believe it. im not saying anything regarding balance of the game, but for fuck sake just stop fucking whining, maybe if P and Z players learned how to play the fucking game instead of spending their entire fucking day writing garbage on teamliquid they would get somewhere. lets look at a few very basic facts that require absolutely no data to be proven
#1 the first game released is from the terran campaign
#2 every fucking post in every god damn starcraft 2 forum says something about OMFG TERRAN IMBALANCED WAT 2 do OMNFG.
these 2 factors alone are going to encourage a ton more people playing terran. more people playing a race is going to lead to a greater development in strategies, leading to a more developed race
Your first point is a reasonable one. Your second one is starting from an assumption and working backwards (assuming that complaints are baseless, which therefore causes unwarranted unpopularity of Zerg), rather than starting from evidence and working to a conclusion.
And if you count the number of units Terran has available, and the number of abilities they have available, you'll see that they simply have more than Zerg, period. Terran isn't better developed because more people play it. More people play it because it is better developed. This is mathematically provable.
|
On September 02 2010 13:14 petzergling wrote:
the dumbest thing by far in this thread is the 1600 and 1500 rating level graph, by far. anyone whos done anything in statistics (or has half a brain) knows that a sample size of 19 (YES THERE ARE 19 PLAYERS ABOVE 1500) is not an adequate sample size to judge something as complex as game balance.
19 is a decent size. Someone really needs to crunch the numbers here. There are statistics you can do, and you can establish uncertainties rigorously even at low sample numbers.
|
I'm SO confused...
Blizzard has already accepted that a balance patch is needed. They have already released information regarding what will be in the balance patch...
Balance patch 1.1 is coming out, it is happening and there is no denying it.
So why are there still players in complete denial?
Game is not perfectly balanced, nerfs are coming, denial is just willful ignorance at this point.
|
On September 02 2010 13:14 petzergling wrote:people in this thread absolutely infuriate me, the terran imbalance bullshit is seriously to a point where i cant believe it. im not saying anything regarding balance of the game, but for fuck sake just stop fucking whining, maybe if P and Z players learned how to play the fucking game instead of spending their entire fucking day writing garbage on teamliquid they would get somewhere. lets look at a few very basic facts that require absolutely no data to be proven #1 the first game released is from the terran campaign #2 every fucking post in every god damn starcraft 2 forum says something about OMFG TERRAN IMBALANCED WAT 2 do OMNFG. these 2 factors alone are going to encourage a ton more people playing terran. more people playing a race is going to lead to a greater development in strategies, leading to a more developed race lets look at some very basic statistics for all of the illiterate 12 year old retards who dont understand how to interpret anything that doesnt tell them UR A GOOD PLAYER U JUST LOST CUS THE GAME IS IMBALANCED number of players at a given level of play does not correlate with balance. AT ALL. lets say fayth made a post in sc2 strategy saying he would give 1000$ to every player who earned 1500 rating as zerg. boy these graphs would look alot different wouldnt they. does that mean that zerg is all of a sudden overpowered and stronger then terran? no. the dumbest thing by far in this thread is the 1600 and 1500 rating level graph, by far. anyone whos done anything in statistics (or has half a brain) knows that a sample size of 19 (YES THERE ARE 19 PLAYERS ABOVE 1500) is not an adequate sample size to judge something as complex as game balance. IM A MORON, WHY IS THIS PETZERGLING? each 1 person on the graph represents over 5% of the total population. lets say a T player with 1501 rating plays a Z player with 1499 rating and the Z player wins (LOL CANT HAPPEN TERRAN OP) this brings an adjustment to our graph of over 10% relative data. considering people at this level play upwards of 10 games a day. proof that this sample size is retarded? here i just re calculated the numbers courtesy of sc2ranks.com and put them on a graph OMFG PROTOSS UPSWING NOW PROTOSS IS IMBALANCED. no the race of players in the top 1500 fluctuates this much on a daily(hourly?) basis k im done see you idiots in disneyland
Why were not banned/warned for this post?
Simply cussing and insulting community members is not a solid argument.
|
I was more willing to buy into this terran IMBA BS before I watched CellaWeRRa's stream... Zergs should watch CellaWeRRa before they determine Terran is too easy. Either CellaWeRRa is the best player in the world period or there simply isn't as huge a divide as numbers/toss/zerg players would lead you to believe. I see Cella beating top terrans all day with zerg, and he makes it look relatively easy (which I know it's not but he does make it look pretty darn easy).
My feeling is that Terran has the easiest learning curve, protoss has the 2nd most difficult, and zerg has the hardest learning curve simply because of micro management needs aka creep tumors, larva, handling more expansions than your opponent at every moment in the game, figuring out a good balance of army to harvester composition with the limited amount of larva you get at a given time.
Since I know a lot of you probably played WoW, I think a fair comparison is to say Terrans are like the Female Night Elf Hunters of SC2 right now. They're just abundant and notorious for easy-mode so having arbitrary graphs and trying to correlate them to race balance is just naive at this stage of SC2s life. Not to mention most self-analyzing casters are Terrans, I often get people asking who streams Protoss or Zerg so they can learn like many do with terran streams but there aren't nearly as many of them.
A few more months and some balance patches will help even everything out I think. I also believe buffing races rather than nerfing powerful races is a better move. Terran is fun to play, nerfing them just nerfs the enjoyment factor. Instead they should buff zerg and make zerg more fun and keep Terran the way it is since it is fun the way it is if they must decide to "balance" it.
|
On September 02 2010 13:27 Eschaton wrote:
Something I'd like to see is the compared win/loss rations per race for random players. Random players should have roughly equivalent experience and skill with each race, and if there is a trend towards random players doing better or worse overall with a race, that would indicate an OP/UP race imo. What made me think of this was taking a look at Danrok's stats page (http://www.danrok.com/stats/) where his win rates for T/P/Z are 56/57/50. I don't think it's possible to pull that info out of sc2ranks.com... maybe that's only info that Blizzard will know.
Sorry for the double post.
While I do sort of agree with you that random players get the whole scope, random in ladders is different than race picking.
I am talking mainly about holding off cheeses, which seems to be so very popular in sc2. Myself as a random player, very rarely have to deal with early cheese shananagins unless its one of "those" types of players (just playing for wins and not improvement).
When I played zerg for a day, i got cheesed more than 50% of the time, as my random zerg, I have ALOT more options for openings.
|
Dear all the terrans that are saying this means nothing (also especially petzergling): It's not just the 1500+ player stats that is overwhelmed with terran. It's from 1200~1500, so sample size of over 300, which obviously isn't set in stone, but also isn't ignorable. If it was just 1500, or just 1400~1500 i would understand the argument, but this is much more wider ranged than people saying "oh it's only 19 people".
|
Take the data and draw your own conclusions?
Mine is very simple:
It doesn't take a fucking graph to realize that terran is just easier to play, has the best units cost for costs and the best synergy among all the possible unit combos You really have to try to make a bad comp as terran that is not gonna give problems to your opponents.
|
PATCH 1.1 IS COMING!
It is a balance patch... It has a ton of nerfs.
Why are people still denying imbalance?!
I just don't understand, please, someone explain.
Blizzard "we have confirmed imbalance, changes are coming."
Players "GAME IS FINE L2P"
I feel like I'm taking crazy pills!
|
On September 02 2010 13:48 Opinion wrote:
PATCH 1.1 IS COMING!
It is a balance patch... It has a ton of nerfs.
Why are people still denying imbalance?!
I just don't understand, please, someone explain.
Blizzard "we have confirmed imbalance, changes are coming."
Players "GAME IS FINE L2P"
I feel like I'm taking crazy pills!
No one's saying it's fine the way it is... but the extreme hyperbole about Terran OPness is starting to wear itself. The upcoming tweaks are IMO good, will they definitely settle the matter? Hell no. We just have to wait it out and let them keep tweaking it.
I'm sure there will be upcoming patches where Terrans get over nerfed, and they get re-buffed. It's almost a certainty.
|
On September 02 2010 07:46 cup of joe wrote:
it means absolutely nothing because the sample size is tiny at the right end of the graph
It does not mean nothing, there are valid arguments for terran being a deserving candidate of nerfs. However, you are correct that given the low sample size at the right end of the graph it denotes more about the players themselves as opposed to the races they are playing.
|
United States22883 Posts
On September 02 2010 13:38 zomgtossrush wrote:Show nested quote +On September 02 2010 13:14 petzergling wrote:people in this thread absolutely infuriate me, the terran imbalance bullshit is seriously to a point where i cant believe it. im not saying anything regarding balance of the game, but for fuck sake just stop fucking whining, maybe if P and Z players learned how to play the fucking game instead of spending their entire fucking day writing garbage on teamliquid they would get somewhere. lets look at a few very basic facts that require absolutely no data to be proven #1 the first game released is from the terran campaign #2 every fucking post in every god damn starcraft 2 forum says something about OMFG TERRAN IMBALANCED WAT 2 do OMNFG. these 2 factors alone are going to encourage a ton more people playing terran. more people playing a race is going to lead to a greater development in strategies, leading to a more developed race lets look at some very basic statistics for all of the illiterate 12 year old retards who dont understand how to interpret anything that doesnt tell them UR A GOOD PLAYER U JUST LOST CUS THE GAME IS IMBALANCED number of players at a given level of play does not correlate with balance. AT ALL. lets say fayth made a post in sc2 strategy saying he would give 1000$ to every player who earned 1500 rating as zerg. boy these graphs would look alot different wouldnt they. does that mean that zerg is all of a sudden overpowered and stronger then terran? no. the dumbest thing by far in this thread is the 1600 and 1500 rating level graph, by far. anyone whos done anything in statistics (or has half a brain) knows that a sample size of 19 (YES THERE ARE 19 PLAYERS ABOVE 1500) is not an adequate sample size to judge something as complex as game balance. IM A MORON, WHY IS THIS PETZERGLING? each 1 person on the graph represents over 5% of the total population. lets say a T player with 1501 rating plays a Z player with 1499 rating and the Z player wins (LOL CANT HAPPEN TERRAN OP) this brings an adjustment to our graph of over 10% relative data. considering people at this level play upwards of 10 games a day. proof that this sample size is retarded? here i just re calculated the numbers courtesy of sc2ranks.com and put them on a graph OMFG PROTOSS UPSWING NOW PROTOSS IS IMBALANCED. no the race of players in the top 1500 fluctuates this much on a daily(hourly?) basis k im done see you idiots in disneyland Why were not banned/warned for this post? Simply cussing and insulting community members is not a solid argument.
Because the community members are absolutely terrible at reading statistics and are wrong. What he lacks in decency and tone, he makes up for by being right. Most of this thread is an example of the worst kind of bean counting there is. Take general, non-nondescript numbers and misinterpret them to mean something very specific.
He's not making a balance argument. He's making a take-Stats-101 argument.
|
On September 02 2010 07:40 ReplayArk wrote:
Shit, we need to nerf Random, they are tearing apart 1500+ diamond.
Also, great fucking post @ petzergling.
+ Show Spoiler +On September 02 2010 13:14 petzergling wrote:people in this thread absolutely infuriate me, the terran imbalance bullshit is seriously to a point where i cant believe it. im not saying anything regarding balance of the game, but for fuck sake just stop fucking whining, maybe if P and Z players learned how to play the fucking game instead of spending their entire fucking day writing garbage on teamliquid they would get somewhere. lets look at a few very basic facts that require absolutely no data to be proven #1 the first game released is from the terran campaign #2 every fucking post in every god damn starcraft 2 forum says something about OMFG TERRAN IMBALANCED WAT 2 do OMNFG. these 2 factors alone are going to encourage a ton more people playing terran. more people playing a race is going to lead to a greater development in strategies, leading to a more developed race lets look at some very basic statistics for all of the illiterate 12 year old retards who dont understand how to interpret anything that doesnt tell them UR A GOOD PLAYER U JUST LOST CUS THE GAME IS IMBALANCED number of players at a given level of play does not correlate with balance. AT ALL. lets say fayth made a post in sc2 strategy saying he would give 1000$ to every player who earned 1500 rating as zerg. boy these graphs would look alot different wouldnt they. does that mean that zerg is all of a sudden overpowered and stronger then terran? no. the dumbest thing by far in this thread is the 1600 and 1500 rating level graph, by far. anyone whos done anything in statistics (or has half a brain) knows that a sample size of 19 (YES THERE ARE 19 PLAYERS ABOVE 1500) is not an adequate sample size to judge something as complex as game balance. IM A MORON, WHY IS THIS PETZERGLING? each 1 person on the graph represents over 5% of the total population. lets say a T player with 1501 rating plays a Z player with 1499 rating and the Z player wins (LOL CANT HAPPEN TERRAN OP) this brings an adjustment to our graph of over 10% relative data. considering people at this level play upwards of 10 games a day. proof that this sample size is retarded? here i just re calculated the numbers courtesy of sc2ranks.com and put them on a graph OMFG PROTOSS UPSWING NOW PROTOSS IS IMBALANCED. no the race of players in the top 1500 fluctuates this much on a daily(hourly?) basis k im done see you idiots in disneyland
|
On September 02 2010 14:02 Jibba wrote:
Because the community members are absolutely terrible at reading statistics and are wrong. What he lacks in decency and tone, he makes up for by being right. Most of this thread is an example of the worst kind of bean counting there is. Take general, non-nondescript numbers and misinterpret them to mean something very specific.
He's not making a balance argument. He's making a take-Stats-101 argument.
Thank you. This isn't a "Terran are OP!" vs "Terran are fine, l2p!" thread. This thread is the perfect example of how dangerous numbers can be in the hands of the uninitiated.
|
On September 02 2010 10:37 ReplayArk wrote:@Cloak, I would like to see the source for this number.
It was in the thread with the 20,000 samples. Protoss had like ~42% prevalence vs. Terran's ~27%. I forgot the name of it, but someone will remember. It was a little dated and done back in early August.
|
On September 02 2010 07:40 ReplayArk wrote:I used some numbers from ( sc2ranks.com) to look at the Racial Distribution at high niveau. My visualisation of the numbers can be found here: While it is known that Blizzard's Matchmaking Algorithm is functional (TL Thread), it was not clear how the same win percentage per race was computed. Some told that there is a racial divergence between diverse skill levels (Bronze League to Diamond League), because new player tend to stay at the campaign's race Terran and the Zerg mechanics may be harder to learn. An alternative to this hypothesis is that people tend to chose race's which let them easier win. If this would be true we should see a race more often if it makes winning simplier. This does not mean that if a race wins often this race would be more easy to play. Some suggested to compare the racial distribution (RD) at high skill levels as indicator for imbalance or more likely as indicator for the lack of exploitable gamestyle. I think this indicator should be seen as a trend on higher skill levels - and more important as it's change, if either new strategies and tactics are discovered, or there are gameplay changes (patches, etc.) implemented. The upcomming patch will change gameplay in special situations quite heavily and we should see a shift in the high level racial distribution over some time (player's Inertia) till an equilibrium is found again (I am not sure if there will be an equilibrium, but it could exist). The time of the shift and the endtime will be good situations to overlook some presumptions we can create right now. I think it would be interesting to do estimations about the post patch use of various units, since it gives us a little time to figure out new concepts before they are established on all servers. This thread could be general discussion about those changes - if it gets more strategy biased it could be moved to the strategy directory. *Edit: Added Sample Count* *Edit: Good post, with things I would like to see happen* *Edit: Data as CSV* *Edit: Added the "plus" to the 1500 Intervall*
Doesn't add up there are 9 toss in top 20 over 1500 and 8 terran how is the terran graph double the toss
Sry your graph is messed all the terran to toss ratios are fucked up but the zerg to rest seems fine
Also why you wasted your time doing this when the site you were on allready does it is a little strange want to make the data swing your way do we : P
http://sc2ranks.com/stats/race/all/1
|
see the problem with this graph, is that people are gonna read it wrong, and complain about terran even more...
|
On September 02 2010 14:22 Radio.active wrote:
see the problem with this graph, is that people are gonna read it wrong, and complain about terran even more...
Its a very bad graph that is not even close to up to date also agian can get the up to date graph from the same site he is listing his data on that is not right.....
Its all very strange
|
i love how people complain about sample size at the top of the graph
if u were looking for sc1 balance, what would u look at? top korean pros, right?
so if 16 the top 20 korean pros were terran do u then complain about sample size? oh there aren't enough top korean pros to get a sample from?
|
This graph is ridiculous I mean... there are like 20 players in the world above 1500 points AND out of those 20 there are 9 protosses, 7 terrans, 3 zergs and 1 random.
I actually took the time to take a look at the sc2ranks link you posted and it prooves you either made a mistake while making the graphic or you manipulated the information.
here I show u picz:
![[image loading]](http://img683.imageshack.us/img683/6386/71096862.png)
|
United Kingdom16710 Posts
i dont get why people have such a hard time reading the stats. no this has nothing to do with 'sample size' since it clearly states that ALL of the diamond players above 600 points across ALL servers have been taken into account. since the parameters have been set, we are looking at the whole population.
while this is very informative, it doesn't necessarily tell us the full story. it could simply be that diamond zerg players aren't enjoying the game as much as their racial counterparts and are just not massing enough games to push through the 1400+ barrier. however you cannot deny that the huge spike in terrans at the top isn't indicative of what everyone is saying about balance. anyway thanks for the interesting stats.
|
Hmmm I guess the stuff I learned in AP Stats isn't so useless after all I'll see what this actually means
|
On September 02 2010 14:31 travis wrote:
i love how people complain about sample size at the top of the graph
if u were looking for sc1 balance, what would u look at? top korean pros, right?
so if 16 the top 20 korean pros were terran do u then complain about sample size? oh there aren't enough top korean pros to get a sample from?
top korean pros != top 20 ladder
One changes with seasons, the other by the hour.
I'll let you figure out which is which.
|
On September 02 2010 13:14 petzergling wrote:
#1 the first game released is from the terran campaign
#2 every fucking post in every god damn starcraft 2 forum says something about OMFG TERRAN IMBALANCED WAT 2 do OMNFG.
these 2 factors alone are going to encourage a ton more people playing terran. more people playing a race is going to lead to a greater development in strategies, leading to a more developed race
I'm not going to reply to this guy's post specifically because he was banned for it, but this argument comes up time and time again and it's completely retarded for the following reason:
Protoss is the most popular race, not Terran.
Only in the top 200 players (that's 200 out of 60,000+ diamond players) do Terrans outnumber Protoss.
|
i'm not good at math so i might be wrong.
it seems to me that a lot of people are taking a subsection of a population and arbitrarily labeling it the general population (which is something i don't think you're supposed to do). i could very easily say that since 1600 is higher than 1500 so they are better players, therefore, i can disregard all players 1599 and lower and have the population of players 1600+. now that i'm only taking the population of players at 1600 terran imbalance isn't as pronounced and you're all wrong. i may be misusing the word sample size but you wouldn't draw a line of best fit with only two data points.
it also seems to me that separating the graph by random 100 point intervals (what's the basis for separating them by 100 points?) makes the "trend" of terran imbalance look extremely obvious until you look at the sample size of each 100 point interval and notice that as terran imbalance gets higher sample size gets smaller. it also skews the bars to hell making it look way worse than it would be if you combine all the data points of 1200+ players. to me it would make more sense to separate by player rank intervals.
as other people said there are also other factors that can skew the data like how many games each player is playing? normally this wouldn't matter if you were looking at the general population as it would tend to iron itself out (i think), however, the "trend" of imbalance is less than 1000 players. i suppose it would be ok then to say that currently in 1200-1600 terrans are over-represented (if they even are) but this is a very limiting conclusion don't you think? it doesn't tell us anything about 1700+ which doesn't (but probably will) exist. that last line probably doesn't make too much sense but you get what i'm saying right?
|
On September 02 2010 14:31 travis wrote:
i love how people complain about sample size at the top of the graph
if u were looking for sc1 balance, what would u look at? top korean pros, right?
so if 16 the top 20 korean pros were terran do u then complain about sample size? oh there aren't enough top korean pros to get a sample from?
Except for one thing. The Blizzard ladder is not a reliable method of measuring "the top players in the world". Almost Every single top person on the ladder is not a top tier tourny player, and one former one has been a hacker.
If you actually think about "Top tier players" in SC2, it actually pans out quite well. Almost all tournaments have very fair racial distribution, once averaged toghether. MLG was predominantly won by toss players, while IEM had a fair distribution of zerg.
If you look beyond that, the data is not more skewed then it was in BW.
http://www.iccup.com/starcraft/content/news/maps_of_the_week_n13_04.html
Look at race stats. plz get perspective newcomers.
petzergling was just temp banned for 2 days by Jibba.
That account was created on 2008-08-21 21:47:27 and had 405 posts.
Reason: There hasn't been a martyr so right since Jesus Christ.
Amen brotha.
|
On September 02 2010 13:14 petzergling wrote:people in this thread absolutely infuriate me, the terran imbalance bullshit is seriously to a point where i cant believe it. im not saying anything regarding balance of the game, but for fuck sake just stop fucking whining, maybe if P and Z players learned how to play the fucking game instead of spending their entire fucking day writing garbage on teamliquid they would get somewhere. lets look at a few very basic facts that require absolutely no data to be proven #1 the first game released is from the terran campaign #2 every fucking post in every god damn starcraft 2 forum says something about OMFG TERRAN IMBALANCED WAT 2 do OMNFG. these 2 factors alone are going to encourage a ton more people playing terran. more people playing a race is going to lead to a greater development in strategies, leading to a more developed race lets look at some very basic statistics for all of the illiterate 12 year old retards who dont understand how to interpret anything that doesnt tell them UR A GOOD PLAYER U JUST LOST CUS THE GAME IS IMBALANCED number of players at a given level of play does not correlate with balance. AT ALL. lets say fayth made a post in sc2 strategy saying he would give 1000$ to every player who earned 1500 rating as zerg. boy these graphs would look alot different wouldnt they. does that mean that zerg is all of a sudden overpowered and stronger then terran? no. the dumbest thing by far in this thread is the 1600 and 1500 rating level graph, by far. anyone whos done anything in statistics (or has half a brain) knows that a sample size of 19 (YES THERE ARE 19 PLAYERS ABOVE 1500) is not an adequate sample size to judge something as complex as game balance. IM A MORON, WHY IS THIS PETZERGLING? each 1 person on the graph represents over 5% of the total population. lets say a T player with 1501 rating plays a Z player with 1499 rating and the Z player wins (LOL CANT HAPPEN TERRAN OP) this brings an adjustment to our graph of over 10% relative data. considering people at this level play upwards of 10 games a day. proof that this sample size is retarded? here i just re calculated the numbers courtesy of sc2ranks.com and put them on a graph OMFG PROTOSS UPSWING NOW PROTOSS IS IMBALANCED. no the race of players in the top 1500 fluctuates this much on a daily(hourly?) basis k im done see you idiots in disneyland User was temp banned for this post.
Thisx1000.
Moreover, irregardless if the new balancing patches coming out, you can't use race distribution in the game to prove balances. This distribution just shows distribution. It does not hint at anything else, though, you can come up with explanations as to why it is like that.
|
On September 02 2010 08:03 Zombee wrote:
---Na, just kidding, but out of 20 games today I spawned friggin 17 times Zerg xD.
The chance of this happening is 1 in 382,323. Consider yourself lucky, (or unlucky). 
|
Sample size is irrelevant because that is the actual data, not a poll or a drawn sample. You guys are retarded lol
User was warned for this post
|
On September 02 2010 15:05 Keap wrote:
Sample size is irrelevant because that is the actual data, not a poll or a drawn sample. You guys are retarded lol
It IS relevant. We aren't trying to get an estimate of the total population. THAT is when it would be irrelevant. We are using the data for 'imbalance'.
Maybe you should learn your basic statistics before you call other people stupid?
|
It's been a while since I've taken statistics, but I do think every group 1500 and below is statistically significant with a slight margin of error.
|
On September 02 2010 13:50 JoKeR[X] wrote:Show nested quote +On September 02 2010 13:48 Opinion wrote:
PATCH 1.1 IS COMING!
It is a balance patch... It has a ton of nerfs.
Why are people still denying imbalance?!
I just don't understand, please, someone explain.
Blizzard "we have confirmed imbalance, changes are coming."
Players "GAME IS FINE L2P"
I feel like I'm taking crazy pills!
No one's saying it's fine the way it is... but the extreme hyperbole about Terran OPness is starting to wear itself. The upcoming tweaks are IMO good, will they definitely settle the matter? Hell no. We just have to wait it out and let them keep tweaking it. I'm sure there will be upcoming patches where Terrans get over nerfed, and they get re-buffed. It's almost a certainty.
They probably will get over-nerfed, or at least it will feel like it, like the roach.
But I wouldn't call this a "tweak". These seem like pretty significant changes IMO.
|
photoshop magic! you can barely tell
![[image loading]](http://img85.imageshack.us/img85/6366/91018474.png)
this is the combined data of all 1100+ sample size of 803.
terran: 331
protoss: 297
zerg: 175
whoops i left out random. whatever you get the point.
what's more interesting is that every 100 point interval the number of users drops by about half each time. is that blizzards system at work? :D
actually random sees a pretty significant drop to .02%
|
Liquid'Tyler trying to teach statistics to some fools! HELLS YES!
|
On September 02 2010 07:57 Mikilatov wrote:
Pretty eye-opening, it seems.
I'm glad that this graph pulls up an interesting point though, Terrans aren't really that overpowered except at high levels in the hands of 1000+ point diamond players.
Or... it could simply mean that the higher level players are the ones who have fully recognized and taken advantage of Terran's strength. It doesn't at all mean that Terran isn't relatively overpowered at the slightly lower levels as well. At the lower levels, what you could be seeing, is Terran moving on through as they move up the ladder getting more points. And that's just talking about the diamond leagues.
So... if you see relatively more Zerg players in the silver league it wouldn't necessarily mean that they were dominating at that level of play so much as it might suggest equivalent players using other races had moved on up to gold.
|
On September 02 2010 08:38 GoBackToGo wrote:Show nested quote +On September 02 2010 08:32 Wr3k wrote:On September 02 2010 07:55 Liquid`Tyler wrote:
As far as I know, sc2ranks is pretty damn comprehensive of Diamond, especially high Diamond (where it seems some people have a problem with the "sample size"). Yeah, the number of people in the 1500+ group is small but that doesn't mean that there is a problem with the sample size. These numbers aren't extrapolated from a small population of the 1500+ Diamond group. These numbers directly represent that group. You would think people would realize this. It makes me cringe when they cry about sample size when ITS THE ENTIRE POPULATION. so lets say im inventing a new pill to cure something, and i test it on a handful of ppl, lets say 5.. then those 5 ppl i gave my pill to make up the whole "population" the pill was ever tested on. so lets say the test showed no adverse reaction for any of the 5 ppl. given that information, would u like to go ahead and try my new pill? btw: t is slightly imbalanced. they are going to fix it.
No, it would mean those 5 people have no adverse reaction. If you're population is those 5 people. What stats you take on the population are representative of the population. Trying to ask that you should take the pill or not based on your findings makes as much sense as taking statistics of the US population and then using it to say that people in china are 50% likely to eat beef.
Seriously, people there is nothing wrong with the data. How you INTERPRET the data up is to you. .
|
I think people here are so concerned about misuse of the words 'sample size' that they're actually missing the point.
Yes, this is the entire population, not a 'sample' however, when the entire population is 20 you can't reliably use the data to prove a point. I think though, that in a population of 200 (e.g. the top 200) you can begin to draw some conclusions from that data.
The question I ask then is this:
Why is there a trend (High Terran %) in this particular subset of people (the top 200) which goes against the general trend of the Diamond population (High Protoss %).
Hypotheses:
1) Racial imbalance.
2) Better players choose Terran.
3) Statistics are skewed by good players switching to Terran due to percieved imbalance.
4) People haven't figured out how to play the other races yet.
5) Terran are easier to play and the overall skill of players hasn't caught up yet to the point where other races can compete at the top level.
6) Top 200 is just randomly mostly Terran for no good reason, aka "sample size" is too small.
These are the most common explanations I've seen. The most plausible by far to me seems to be the first option, since there's little evidence to support 2 or 3 (except for a couple of notable publicised cases), 4 seems unlikely given there doesn't seem any reason why the rate of development of strategies for one race should be higher than any of the others. 5 is plausible but isn't this also a form of imbalance? 6 is probably true for the top 20, but the top 200? I think this is a big enough number of people to make at least some conclusions.
I think anyone with a brain will agree that imbalance, while not the only plausible explanation, is definitely the most likely.
|
On September 02 2010 08:08 Doomrok wrote:Terran is perfectly balanced, I mean, look at my Terran vs Zerg win rate! http://www.danrok.com/stats/r r:r
Actually... all this really reveals is that you suck at mirror matches. ;P
Love the game notes though!
|
one thing we do need is a time frame... another graph after a month or so.
I feel Terran VS P or Z could be more balance but its more of a sense that i have from playing too many games. Nothing i could substantiate.
This chart is good, another chart after a month or so, it would help make this more credible since technically although this is a large sample size, its of one moment in time.
Its probably a LOAD of work but to see the progress of race distribution amoungst mid/high diamonds monthly or so.
*blizzard should look at this thread!
|
On September 02 2010 11:44 roronoe wrote:
I'm interested in why random starts off high, dips in the middle, and goes back up, in a really consistent manner too.
Anyone have a plausible explanation for this?
Low level diamonds can cheese pretty well, random in lower diamond means that you can't do race specific cheeses // aggressive plays until you scout.
In the mid level, diamond players dont rely on cheese as much, and have more familiarity with their race then whatever random player has, IOW their stock standard game is better. Even though the random player knows what race the opponent is, they might not necessarily use a tailored build vs that race (which might give them an advantage) because they might be preparing for cheese as they are used to.
Then as they gain ranking // skill again, randoms have learned how to scout // deal with cheese well enough such that they can use tailored builds vs opponents race or are familiar enough with all the races to beat the opponent in a macro game.
|
nevermind what the hell am i saying
|
On September 02 2010 15:48 mahnini wrote:Show nested quote +On September 02 2010 15:40 Wargizmo wrote:+ Show Spoiler +
I think people here are so concerned about misuse of the words 'sample size' that they're actually missing the point.
Yes, this is the entire population, not a 'sample' however, when the entire population is 20 you can't reliably use the data to prove a point. I think though, that in a population of 200 (e.g. the top 200) you can begin to draw some conclusions from that data.
The question I ask then is this:
Why is there a trend (High Terran %) in this particular subset of people (the top 200) which goes against the general trend of the Diamond population (High Protoss %).
Hypotheses:
1) Racial imbalance.
2) Better players choose Terran.
3) Statistics are skewed by good players switching to Terran due to percieved imbalance.
4) People haven't figured out how to play the other races yet.
5) Terran are easier to play and the overall skill of players hasn't caught up yet to the point where other races can compete at the top level.
6) Top 200 is just randomly mostly Terran for no good reason, aka "sample size" is too small.
These are the most common explanations I've seen. The most plausible by far to me seems to be the first option, since there's little evidence to support 2 or 3 (except for a couple of notable publicised cases), 4 seems unlikely given there doesn't seem any reason why the rate of development of strategies for one race should be higher than any of the others. 5 is plausible but isn't this also a form of imbalance? 6 is probably true for the top 20, but the top 200? I think this is a big enough number of people to make at least some conclusions.
I think anyone with a brain will agree that imbalance, while not the only plausible explanation, is definitely the most likely.
if you were to accept that racial imbalance causes terrans to be placed higher why is the trend only happening at the highest ELOs? how come at lower ELOs dont see any terran inflation? shouldn't racial imbalance cause a larger discrepancy causing an upside down pyramid of terran distribution? i know upside down pyramid is stupid i dont know how else to say that right now ~_~
This was explained by a few people earlier in the thread.
It's all due to the matchmaking system trying to get everyone to have a ~50% win loss ratio, top players who play underpowered race(s) will be placed in lower brackets and make the distribution seem even in those brackets.
Basically if there's an imbalance, the only place it will ever show up is at the top since at every other level there will still be loads of people who use the 'underpowered' race(s) because they're players who should be rated higher but have been pushed down due to the imbalance.
Edit: I guess I was too quick for your edit but I'm going to leave my post up because I think a lot of people don't seem to understand this.
|
1: Anyone using the phrase "sample size" in an attempt to debunk the theory that Terran are imbalanced, immediately invalidates their own argument......there is no "sample size".
The data is an actual representation of the entire population, not a sample or a subset.
2: If it is accurate.(no corroborating source or datum is evident) then my interpretation
is that Terran are Imba! to argue otherwise makes it sound like you have your head in the sand.
on its own the data is not so damning ..but in concert with the overwhelming amount of anecdotal evidence surely it provides clear and incisive view off the issue.
there is no forum anywhere in the world where i have seen the claim that Zerg are to strong
follow by masses of butt-sore Zergs saying...no other races just need to learn how to play
yet we see this every day about Terran(why do you think this is so?)
what is more likely; top Terran's are just overwhelmingly better than top Zergs ....OR....
There is a mismatch regarding the effectiveness/supply cost/ease of use of Terran units that only becomes apparent when the issue of skill is not in variance (i.e. the very top)?
|
Looks like random needs a buff.
|
1: Anyone using the phrase "sample size" in an attempt to debunk the theory that Terran are imbalanced, immediately invalidates their own argument......there is no "sample size".
The data is an actual representation of the entire population, not a sample or a subset.
If there were only 4 starcraft 2 players in the entire world, and you took data on the entire population, your sample size is 4. It doesn't matter that it's the entire population, it's still a very small sample size.
While the data at 600, 700, etc points seems robust, the sample size of 1500+ players is laughably small to do any statistics on. It doesn't matter that he used every 1500+ player reported; there are simply too few 1500+ players in the world to get meaningful data about.
It's also entirely unproven that people pick the race they feel is strongest. Some players may pick their race more for style, or on a whim. Others may try to pick the strongest race, but be wrong about it, and actually pick a weaker race. Granted the top of the top are less likely to be mistaken, but I'd say it's still pretty common.
|
I dont like these useless statistics. You need so much more data to make this usefull.
How many players play the different races?
How many players are even 1400+?
You cant just pick the top 5 and go, "well there are 3 terrans so terran is OP". You need info across all brackets, not just the top 2% or whatever.
I'd like for someone to do that, unfortunately i'm just too lazy :D
|
I would like to see a graph of each race over time.
I have seen on these sites that keep track of each race that zerg popularity overall has gone down from ~25% to 23% in the past month. Where are all the Zergies going?
|
Okay, guys, seriously. I hate all of you right now.
On one hand, people shouting that Terran isn't OP are clearly wrong. Blizzard agrees. Done.
On the other hand, people shouting that Terran is clearly OP because of this graph are also wrong. I haven't seen this rebuttal brought up, so forgive me if I'm repeating someone.
There seems to be an underlying assumption that the proper ratio of players across the races (ie, the null hypothesis) should be some variation on 30-30-30-10, with the 10 being those overachieving randoms out there.
Unfortunately, this is a faulty assumption. Fortunately, we have a means of examining this.
Because of Blizzard's stunningly effective matchmaking algorithm, players maintain a 50-50 win ratio except at the extreme high and the extreme low end. A positive win ratio will result in promotion to a higher league (and tougher players), and a negative win ratio will result in the reverse. This has the effect of pushing extremely skilled and unskilled players to the edges of the bell curve -- high diamond and gutter bronze.
Examining the edges of the bell curve is an extremely poor way of judging the performance of the remainder of the curve due to the fact that individual skill differences and irrational race preferences become more and more pronounced as the group diminishes in size. In other words, the power of a statistical analysis is that over a large enough group size, individual biases and skill differentials will largely cancel each other out and leave pure, unadulterated signal behind -- within a measure of confidence, of course. At the very top and very bottom this can no longer be counted upon, as irrational individual choices and the fundamental fact that some players are just plain better than others has a huge impact on what appears to be a rational decision (ie, which race will I play?).
To put it even more simply -- I have seen the claim that because Terran dominates the extremely small high-end skill bracket that their race is overpowered. This claim is based on the assumption of even skill spread across all races, and even racial representation across all players, all in a group of 50-odd individuals.
That is a remarkably strong claim for a group size so small and I find no reason to believe it is true. When talking about slices of 5 or 10 players, why am I to believe that the Terrans are not simply better Starcraft 2 players? Or that there is not some natural human preference towards Terran? It is certainly not as preposterous as some posters here have implied.
Now, lets see if we can come up with a more accurate null hypothesis, as I think the 30-30-30-10 assumption is pretty clearly flawed. Take a look at the middle of the bell curve -- Gold League.
While I know Gold League represents something like the 50-70% percentile and is thus somewhat off-center in terms of skill, it is the league located farthest away from the extremes (diamond and bronze) and so the effects of very, very good and very, very bad players can be ignored. These are the solid middle children of the laddering family, and from them we can derive a proper 'null' hypothesis with which to compare those above and below.
SC2Ranks reports the following. In Gold League,
38% play Protoss
32% play Terran
20% play Zerg
10% play Random
I see little reason not to adopt this spread as a proper null hypothesis of initial racial division, reflecting all the biases of gamers who play Starcraft 2 without all the high-end super-refinement concerns of the high diamond league or the unstable nuggets of anti-information on which Bronze is fueled. Due to their skill limitations or build misconceptions, these Gold players have not been worthy of a promotion to higher leagues, nor have they been demoted. They are the average -- and note, if you add in the other two 'middle' leagues, Platinum and Silver, the percentages are largely unchanged while near-tripling the sample size. This is the basis from which we should evaluate the racial spread of the higher and lower leagues.
Now lets look at the data from the graph at the 1100 bracket, which is the last data point with a group size I feel comfortable examining. It shows:
41% play Protoss
35% play Terran
23% play Zerg
1% play Random
Compared to the neutral Gold League numbers, only Random suffers markedly. Players appear to have largely settled on their races, and each race has benefited equally.
This says to me that there is no clear preference amongst the races compared to the null hypothesis. Therefore, looking at the lack of Zerg and concluding that they are underpowered does not follow. Other reasons besides win percentage determine racial preference, as shown by the leagues where win percentage is 50-50, by definition.
tl;dr -- Examining the extreme high and low end of any bell curve is unhelpful as differences in player skill and player preference skew what should be impartial data. Also, the assumption of a natural even split of players amongst all races finds no support in the data, therefore few solid conclusions can be drawn from the provided graph.
|
You forgot to include total number of players for each sample.
|
wow this kinda shows something about terran
ps. i did a statisticks corse BIATCH
|
On September 02 2010 17:33 Deadlyfish wrote:
I dont like these useless statistics. You need so much more data to make this usefull.
How many players play the different races?
How many players are even 1400+?
You cant just pick the top 5 and go, "well there are 3 terrans so terran is OP". You need info across all brackets, not just the top 2% or whatever.
I'd like for someone to do that, unfortunately i'm just too lazy :D
These statistics are not useless at all. First of all the number of players is written underneath each column. And somebody in this thread (to lazy to search) already did the statistics on the data. You would expect 33% to be terran, if the races were randomly distributed. If your sample size is small, you need a higher divergence from these 33% to be significant and statistic tests factor that already in. In the highest braket you got 60% (!!!) terran out of 20 players. Lower ratio in the lower brackets, however you got a larger amounts of players making this maybe even more significant.
In conclusion: There is no doubt, that the distribution of terrans is not random at the top.
Why so many terrans are at the top everyone can answer for himself. Maybe it is imbalance, maybe it is because zergs don't know to play their race aka "abuse mobility, morph 60 food counter army in seconds and nydus ultras into enemy mineral line". Chose what ever reason you like. :D
|
On September 02 2010 17:33 Deadlyfish wrote:
I dont like these useless statistics. You need so much more data to make this usefull.
How many players play the different races?
How many players are even 1400+?
You cant just pick the top 5 and go, "well there are 3 terrans so terran is OP". You need info across all brackets, not just the top 2% or whatever.
I'd like for someone to do that, unfortunately i'm just too lazy :D
I'm pretty sick of reading posts like this in the thread as it's been explained: The system is basically keeping potentially good zerg/protoss users down at the lower levels making the groups seem more evenly distributed when the reality is that the better players of those races are being artificially held down by imbalance.
Just scrolling through the top 200 reveals that wayyy more terrans have 66% wins or higher(which suggest upward mobility is possible for the player) so as the ladder goes on at the top this will just get worse as Zerg/Protoss users just grind out points on vs P/Z matches and continually lose to Terrans.
|
On September 02 2010 08:21 MamiyaOtaru wrote:
I don't know how anyone can read this as anything but evidence that Terran is OP. Automatic match making keeps everyone at around %50 win ratio everywhere but the very top. The only way to tell balance with AMM around is by looking at the very top, and that's what this graph does. Terran is way overrepresented at the top, hence it is OP.
And I can't agree with people saying that the graph only shows imbalance at the top so it's fine everywhere. As has been said many times the top is the only place it *can* be showed with AMM in the mix. And it's there.
And for all the zerg players lower in diamond, this graph doesn't show that it's balanced for them at their level. They might be at a higher level (increasing their representation on the right side of tthe graph) if they weren't underpowered. Same for zerg in silver. They are getting %50 wins thanks to AMM but some of the higher ones might be in gold if it were more balanced.
With an even distribution of skill this isn't going to show until you look at the very top. This graph does, and the result is there plain as day.
This point needs to be repeated until everyone understands it. Thanks for explaining it in a clear way.
|
excellent graph , couldn't be clearer
bring on patch 1.1 (and more terran nerf patches)
|
|
|
On September 02 2010 17:48 Winter_mute wrote:Show nested quote +On September 02 2010 17:33 Deadlyfish wrote:
I dont like these useless statistics. You need so much more data to make this usefull.
How many players play the different races?
How many players are even 1400+?
You cant just pick the top 5 and go, "well there are 3 terrans so terran is OP". You need info across all brackets, not just the top 2% or whatever.
I'd like for someone to do that, unfortunately i'm just too lazy :D These statistics are not useless at all. First of all the number of players is written underneath each column. And somebody in this thread (to lazy to search) already did the statistics on the data. You would expect 33% to be terran, if the races were randomly distributed. If your sample size is small, you need a higher divergence from these 33% to be significant and statistic tests factor that already in. In the highest braket you got 60% (!!!) terran out of 20 players. Lower ratio in the lower brackets, however you got a larger amounts of players making this maybe even more significant. In conclusion: There is no doubt, that the distribution of terrans is not random at the top. Why so many terrans are at the top everyone can answer for himself. Maybe it is imbalance, maybe it is because zergs don't know to play their race aka "abuse mobility, morph 60 food counter army in seconds and nydus ultras into enemy mineral line". Chose what ever reason you like. :D
20 players doesnt mean anything. Thats why the 1400+ and 1500+ data is useless. Would be better to do 1200+ as the max because atleast then you'd have more than 20 people.
And no, you wouldnt expect 33% to be terran since there are way more terrans than zergs overall. Also, you cant just pick the top 1 or 2% and stop there. You have to take races across the divisions to get a clear picture. Lets say there were 100% protoss in plat, but 60% terran in diamond. You wouldnt get a clear picture by just using diamond as your sample would you?
I agree that Terran is too good, and i like the new patch, but people need to relax and realise that there isnt as much unbalance as they think. I hope that everyone agrees that Terran is too good and zerg is abit weak, but that doesnt make this graph correct.
|
On September 02 2010 08:21 MamiyaOtaru wrote:
I don't know how anyone can read this as anything but evidence that Terran is OP. Automatic match making keeps everyone at around %50 win ratio everywhere but the very top. The only way to tell balance with AMM around is by looking at the very top, and that's what this graph does. Terran is way overrepresented at the top, hence it is OP.
And I can't agree with people saying that the graph only shows imbalance at the top so it's fine everywhere. As has been said many times the top is the only place it *can* be showed with AMM in the mix. And it's there.
And for all the zerg players lower in diamond, this graph doesn't show that it's balanced for them at their level. They might be at a higher level (increasing their representation on the right side of tthe graph) if they weren't underpowered. Same for zerg in silver. They are getting %50 wins thanks to AMM but some of the higher ones might be in gold if it were more balanced.
With an even distribution of skill this isn't going to show until you look at the very top. This graph does, and the result is there plain as day.
I will quote this one too.
One example of this was watching Morrow vs Dimaga play yesterday in ESL, and as a long time fan of broodwar/warcraft 3/etc it just made my sad how Dimaga was playing at least one level above Morrow and still lost.
|
Why would top players choose Terran more than other races? Either because they just liked them better, or because they find them to have an advantage. Now, the real question would be to compare the Diamond league player's races to the race they played in BW. If someone was awesome at BroodWar with Zerg, why would they change unless they felt there was an advantage or weakness on one side or the other?
Just an idea for another way to aggregate the numbers to get some stats and results that could be interesting.
|
Actually, I found some numbers I pulled off rts-sanctuary on august 13th and I just pulled the same set of info right now.
Its kinda funny looking at the top 20, where the smallest domination is terran in the US with 9, there's 13 terran in the top 20 for europe and even korea went from 7 to 10.
Top zergs have gone down from 5 to 3 world wide, and have taken considerable drops at all levels, even in zerg-run korea zerg went from 8 to 5 in the top 20.
Across the board, Protoss hasn't really moved except in north america, there's 9 top protoss there, up from 7 (only 1 zerg in the US top 20).
Meanwhile, Terran has become more popular among all diamond players while both zerg and protoss becomes less.
The game really is moving towards TvT all day.
I can post my numbers if somebody wants to turn them into graphs.
|
On September 02 2010 18:05 Deadlyfish wrote:Show nested quote +On September 02 2010 17:48 Winter_mute wrote:On September 02 2010 17:33 Deadlyfish wrote:
I dont like these useless statistics. You need so much more data to make this usefull.
How many players play the different races?
How many players are even 1400+?
You cant just pick the top 5 and go, "well there are 3 terrans so terran is OP". You need info across all brackets, not just the top 2% or whatever.
I'd like for someone to do that, unfortunately i'm just too lazy :D These statistics are not useless at all. First of all the number of players is written underneath each column. And somebody in this thread (to lazy to search) already did the statistics on the data. You would expect 33% to be terran, if the races were randomly distributed. If your sample size is small, you need a higher divergence from these 33% to be significant and statistic tests factor that already in. In the highest braket you got 60% (!!!) terran out of 20 players. Lower ratio in the lower brackets, however you got a larger amounts of players making this maybe even more significant. In conclusion: There is no doubt, that the distribution of terrans is not random at the top. Why so many terrans are at the top everyone can answer for himself. Maybe it is imbalance, maybe it is because zergs don't know to play their race aka "abuse mobility, morph 60 food counter army in seconds and nydus ultras into enemy mineral line". Chose what ever reason you like. :D 20 players doesnt mean anything. Thats why the 1400+ and 1500+ data is useless. Would be better to do 1200+ as the max because atleast then you'd have more than 20 people. And no, you wouldnt expect 33% to be terran since there are way more terrans than zergs overall. Also, you cant just pick the top 1 or 2% and stop there. You have to take races across the divisions to get a clear picture. Lets say there were 100% protoss in plat, but 60% terran in diamond. You wouldnt get a clear picture by just using diamond as your sample would you? I agree that Terran is too good, and i like the new patch, but people need to relax and realise that there isnt as much unbalance as they think. I hope that everyone agrees that Terran is too good and zerg is abit weak, but that doesnt make this graph correct.
20 players does mean something. Again: statistic tests already take the amount of players into acount. It is not that we randomly pick 20 people out of a gigantic pool, we specifically pick the top players.
And yes, you would expect 33% to be terran. Yes there are way more terrans than zerg (in fact more terran than there should be, if it were a random distribution). But even if you normalize the data to the overall percentage of terran players, because you think people just choose terran for reasons of familiarity or design, you still get a significant divergence from the expected outcome. You got roughly 25% zerg
players and in the top bracket there are only 10%.
Also, you HAVE to pick the top x% to see this imbalance => see match making system. (In fact we are already picking the top 10% (??) of all players, because we only look at diamond.) If all terran players were shifted by, lets say 50-100 points towards the top, then you would only see this in the top and in the bottom.
And regarding your example, if you just look at diamond you will get a clear picture ... for diamond. This data is for diamond players, noone said it was valid for non-diamond players.
Edit: typo
|
Must feel good to be a Zerg player at the top.
|
On September 02 2010 18:05 Deadlyfish wrote:20 players doesnt mean anything. Thats why the 1400+ and 1500+ data is useless. Would be better to do 1200+ as the max because atleast then you'd have more than 20 people. And no, you wouldnt expect 33% to be terran since there are way more terrans than zergs overall. Also, you cant just pick the top 1 or 2% and stop there. You have to take races across the divisions to get a clear picture. Lets say there were 100% protoss in plat, but 60% terran in diamond. You wouldnt get a clear picture by just using diamond as your sample would you? I agree that Terran is too good, and i like the new patch, but people need to relax and realise that there isnt as much unbalance as they think. I hope that everyone agrees that Terran is too good and zerg is abit weak, but that doesnt make this graph correct.
It has been said numerous times already in this thread, but I guess I'll summarize it for you:
A. The matchmaking engine equalizes everybody expect the top, because you are matched with better players. Example: If you're in bronze and win 5 times in a row you will be paired with silver players. once you start beating them you get matched with gold. The top of diamond doesn't have this mechanism so the balance between races become more clear there.
B. THIS ISN'T A SAMPLE, it is the observation of reality, the whole population so to speak. A sample in statistics is a fraction of the population, and then you indeed have to adhere to certain rules like random distribution to be able to say that the sample indeed reflects the population as a whole. Since this isn't a sample, it basically isn't statistics to begin with, just an observation of reality. So your statement about 20 players not meaning anything is false, they mean everything, since they are everything...
on topic: I checked this out on sc2ranks too but back when the max was 1300+. The race distributions are still more or less the same. It would be interesting to see the development over time around patches to see how they affect the top of the league
|
what I read from this is:
1. protoss does 4gate pressure and wins (almost) regardless of the protoss' skill level, the 4gate can only be done/improved so much and it caps skill roof for protoss very early. higher up it doesn't work and there goes the entire protoss gameplay out the window
protoss is truly a terribly designed race atm
2. terran is very strong
3. zerg is very frustrating to play
good post by OP i must say btw
|
On September 02 2010 17:24 Tasonir wrote:Show nested quote +1: Anyone using the phrase "sample size" in an attempt to debunk the theory that Terran are imbalanced, immediately invalidates their own argument......there is no "sample size".
The data is an actual representation of the entire population, not a sample or a subset. If there were only 4 starcraft 2 players in the entire world, and you took data on the entire population, your sample size is 4. It doesn't matter that it's the entire population, it's still a very small sample size. While the data at 600, 700, etc points seems robust, the sample size of 1500+ players is laughably small to do any statistics on. It doesn't matter that he used every 1500+ player reported; there are simply too few 1500+ players in the world to get meaningful data about. It's also entirely unproven that people pick the race they feel is strongest. Some players may pick their race more for style, or on a whim. Others may try to pick the strongest race, but be wrong about it, and actually pick a weaker race. Granted the top of the top are less likely to be mistaken, but I'd say it's still pretty common.
dude...
did you just say there is no valuable information to be interpreted from these numbers?
IF there were only 4 SC 2 players in the world, which there is not. (why start supposing hypothetical's to prove a point that is evident in the actual situation)
data in regards to the race selection and performance would still be of some use for balance discussion...
its like there are Terran police officers in this thread telling us to "move along, Nothing to see here"
your last paragraph completely misses the point..i am invoking Occam's Razor to best explain why we are seeing this data and these anecdotal evidence trends
I'm starting to know how Copernicus must have felt.........."guys, guys... the earth actually revolves around the sun"< " not the other way around".
"Dont be stupid Copernicus"
"but guys the evidence clearly points to this being a fact and it explains a lot of stuff"
"stop whining Copernicus your just a baby, shut up take your burning at the stake"
|
On September 02 2010 07:46 cup of joe wrote:
it means absolutely nothing because the sample size is tiny at the right end of the graph
Uhhhh sample size? That isn't a sample size. That's POPULATION. If you have 10 people in the world you don't take a sample of 9, you look at all 10.
EDIT: If population is available, you don't take samples. Samples are representative of the population. Here the population is small enough that taking a smaple isn't warranted.
|
I hate be beating this dead horse more, but how can people say that the 'sample size' at the top does not matter? (nothing wrong with using the word samplesize even though its.the entire top population.) Now Im not saying.this chart is wrong or that terran is perfectly balanced or whatever but looking at the top 60 it looks like terran suddenly gains a HUGE edge. But as far as we know 20 terrans might be at 1401 points and 20 z at 1399 points then 1 day later all t lose 1 game and all z win one and the graph would look hugely different. It's like If there was only 1 player above 1500 and that was say a p ( I know both huk and ttone have been top rated lots) then 1500+ would just be a huge yellow bar. Would people still say sample doesn't matter and that p is obviously op at top level play?
|
On September 02 2010 19:28 Earll wrote:
I hate be beating this dead horse more, but how can people say that the 'sample size' at the top does not matter? (nothing wrong with using the word samplesize even though its.the entire top population.) Now Im not saying.this chart is wrong or that terran is perfectly balanced or whatever but looking at the top 60 it looks like terran suddenly gains a HUGE edge. But as far as we know 20 terrans might be at 1401 points and 20 z at 1399 points then 1 day later all t lose 1 game and all z win one and the graph would look hugely different. It's like If there was only 1 player above 1500 and that was say a p ( I know both huk and ttone have been top rated lots) then 1500+ would just be a huge yellow bar. Would people still say sample doesn't matter and that p is obviously op at top level play?
A dead horse indeed. See:
On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200.
|
Anyone else read this thread title and expect the options to be Asian, Caucasian, African, etc.?
|
All I see is zerg definitely not getting played.. Anybody have stats of when roaches only needed 1 supply?
|
On September 02 2010 19:48 makopluxx wrote:
Anyone else read this thread title and expect the options to be Asian, Caucasian, African, etc.?
errm no?
|
On September 02 2010 17:57 iPlaY.NettleS wrote:
excellent graph , couldn't be clearer
bring on patch 1.1 (and more terran nerf patches)
Unfortunately, this is also going to bring on unnecessary Protoss changes as well. : (
|
On September06 Modular1 wrote:Show nested quote +On September 02 2010 17:57 iPlaY.NettleS wrote:
excellent graph , couldn't be clearer
bring on patch 1.1 (and more terran nerf patches) Unfortunately, this is also going to bring on unnecessary Protoss changes as well. : (
not to mention the changes to terran are not good....
siege tank nerf is a mistake, what they should have done was change the splash damage mechanic back to how it was in the start of beta (splash damage originates from impact site rather than target unit), this change would have the effect of making zerglings and upgraded zealots good against tanks again..
BC change was just wtf out of the blue
zealot change will have to wait and see, possibly leaves protoss volnurable to early pool rush (yet another side effect from blizzards crap&tiny maps causing problems in gameplay)
reaper change just makes people not go reapers, removing depth from gameplay
bunker change was not what was needed, they needed to change salvage not build time...
|
Really good graph, it doesn't say everything but even at a glance it's obvious that races are not even. The discussion in this thread disappoints me though, I dunno if TL was better before SC2 but the signal/noise ratio in this thread is depressing. Whenever balance is involved kids get up on their white horses and batter the opposition with blunt reasoning, it's not what you'd expect from strategists.
The hidden factor here is skill. We should assume that skill is evenly distributed between races. This assumption might be false, but without it there's nothing to discuss. We cannot measure skill directly. Points are calculated from wins and losses and are an attempt to measure skill. If skill is equally distributed between races then points should be evenly distributed between races.
Looking at the race distribution in the graph we see that the ratios are skewed at the top. The points are not evenly distributed between races. That means race is a factor that affects player points. Points are earned from winning and lost from losing. This means race is a factor that affects whether players win or lose. That's another way to say racial imbalance.
Our sample is the entire population of diamond players, the probability that the graph is representative of that population is 100%. (And just to clarify: that probability wouldn't be noticeably lower if the sample was 1/10th of the population).
Blizzard cares about balance in their games. They do take their time though, I suppose it's the downside of their quality focused stance to development. Even though the game is still young and tactics are evolving I'd love to see somewhat faster patching. Some of their announced changes for 1.1 are not huge but they would improve the current situation. For SC2 as an e-sport imbalance is a big threat, there's big money and even careers on the line. Making faster smaller changes would show the world that Blizzard is on the ball about balancing SC2. That said I'm still confident that Blizzard will make SC2 an even better game as time goes by, it just feels slow at times.
|
Terrans are just the better players! Why cant you understand it??
-______________-
Srsly, everyone knows T's imba, esp. vs Z. Everyone who denies it, plays terran himself or is stupid -,-
|
This graph doesn't, by itself, mean that Terran is better than the rest. It means Terran is simply more popular among the top. To actually judge racial parity, you'd need to do something like this: keep a win/loss record of all 1000+ diamond players' games and what race they played on each game, then find the total win ratio for Terran, Protoss, and Zerg among all of those games. This will tell you how well those races fair among the top--not how popular they are.
|
On September 02 2010 20:56 carwashguy wrote:
This graph doesn't, by itself, mean that Terran is better than the rest. It means Terran is simply more popular among the top. To actually judge racial parity, you'd need to do something like this: keep a win/loss record of all 1000+ diamond players' games and what race they played on each game, then find the total win ratio for Terran, Protoss, and Zerg among all of those games. This will tell you how well those races fair among the top--not how popular they are.
It sure doesn't mean that way, in the sense that it makes certain that notion, but it does a fair bit to suggest so. It strongly supports the hypothesis that it is easier to get to the top by playing terran, and harder when you play zerg. Again, it doesn't prove in a definite sense that conclusion, but it supports it.
|
On September 02 2010 20:56 IPS.Mardow. wrote:
Terrans are just the better players! Why cant you understand it??
-______________-
Srsly, everyone knows T's imba, esp. vs Z. Everyone who denies it, plays terran himself or is stupid -,-
Yeah, well-said. I really don't have a problem with people playing terran, what I have a problem with is people who play Terran who blatantly denies any imbalance, and tries their best to explain away any statistics and evidence that suggests so.
|
Another point is, since there is few zergs on top, who plays alot more ZvT then the terrans play TvZ, then shouldnt the zergs have more practice and be better in that matchup?
|
The biggest problem I have with the sc2ranks.com stats is, that they lack a global overall average.
That global average represents the race popularity for each race. In different regions it represents the popularity in the regions.
But when split by league/points, it doesn't represent popularity in the leagues. Because you end up in different leagues by winning or losing, not just by picking your race.
The most interesting numbers regarding balance are, how the race distribution diverges from the overall average in the different leagues.
Now we can say, there a re less terrans than average in the lower ranks, and more terrans than average in the upper ranks. This has nothing to do with popularity. Because in a balanced game, every league would have abot the same distribution as the overall average. In an imbalanced game the different winning percentages for the races skew the numbers. And in exactly the way that can be observed: higher percentage of that race in the uper ranks, lower percentage of that race in the lower ranks.
|
How much player are 1400+? or even 1500+? This only means that there are more high skilled terrans then other races. Graphs like this say nothing about balance
I accept that Terran needs some Balance changes, but please dont take this as a free ticket to qq when you are losing. The imbalance is very small on the skill level of most of the players here around, and most of the time when you lose to terra, it will be some mistake u made.
|
On September 02 2010 21:40 Daxten wrote:
How much player are 1400+? or even 1500+? This only means that there are more high skilled terrans then other races. Graphs like this say nothing about balance
I accept that Terran needs some Balance changes, but please dont take this as a free ticket to qq when you are losing. The imbalance is very small and on the skill level of most of the players here around, and most of the time you lose to terra it will be some mistake u made
It can also mean it is easier to get to the top with terran, and hard with zerg. It can also mean players are switching to terran because it is simply easier to win with. imbalance is very small and only at the top? Sorry but you are quite wrong in that regard. ZvT is a lot harder than TvZ, for reasons I shall not name for fear of sidetracking this thread, but that is fact and is recognisable from middle platinum onward.
*Sorry I know this is not the place to discuss this in depth but it is the sort of terran attitude that contributes nothing to a statistical discussion
|
very sad statistics. i hope there is more balance to make the distribution more equal. a competitive game like Starcraft should not see such lopsided pickings.
|
On September 02 2010 09:07 imbecile wrote:
1. If you win a lot, you are removed from the lower rankings and promoted to the higher rankings.
2. To be in the highest rankings, you must consistently win a lot.
3. In the highest ranking tiers, there are more terrans than all other three races combined
4. The percentages also differ greatly from all other ranking tiers and from the overall average in favor of terran.
So those tiered percentages not only represent a current state, but a development, a history.
There are only 2 explanations for the discrepancies:
a) the best players like playing terran a lot more than all other players
b) Playing terran gives you consistently more wins compared to the other races, especially at the highest level.
And it seems to be consensus, that everyone below 1000 points is more or less a casual player.
And it also should be noted that if the lower ranks have balanced winning percentages per race, it doesn't mean that much. Because those that benefit from imbalances quickly disappear from the lower ranks. The process of promotion and demotion enforces balanced outcomes in the lower ranks. The discrepancies are passed upwards, until there is no upwards anymore where it can be passed to, and there they start to accumulate.
Your username is very very misleading :p . I've been saying something along these lines for some time now and still people refuse to understand it.
Balance for the sake of fair gameplay is not even my biggest concern regarding this game, I'm actually more concerned with insufficient skill caps, lack of innovative play due to poor race design, map hackers and GOMtv charging too much!
|
I actually took the time to take a look at the sc2ranks link you posted and it prooves you either made a mistake while making the graphic or you manipulated the information.
@West or you simple forgot that the time I posted it there was another racial distribution that is now. You also got the CSV with the data at the time I took the values from sc2ranks.com If you compare it with another time and look at the differences you are doing what I suggested in my post, still if you do the math your pic looks quite the same as my values are.
Furthermore I want to highlight again, if you measure things at different times you could get different results, since something could increase, for example the distribution of really large NBA player's is way different than fourthy years ago.
You forgot to include total number of players for each sample.
@Jarl No, I don't know why I should include it to each sample. The data is in the post, if you want to further add the total numbers of players for each sample (which is partly written above the point intervall) you are free to do so, the CSV data is available and I like it if people check my results. But if you think something is missing you are free to make graphics better and findings more solid.
|
On September 02 2010 13:14 petzergling wrote:people in this thread absolutely infuriate me, the terran imbalance bullshit is seriously to a point where i cant believe it. im not saying anything regarding balance of the game, but for fuck sake just stop fucking whining, maybe if P and Z players learned how to play the fucking game instead of spending their entire fucking day writing garbage on teamliquid they would get somewhere. lets look at a few very basic facts that require absolutely no data to be proven #1 the first game released is from the terran campaign #2 every fucking post in every god damn starcraft 2 forum says something about OMFG TERRAN IMBALANCED WAT 2 do OMNFG. these 2 factors alone are going to encourage a ton more people playing terran. more people playing a race is going to lead to a greater development in strategies, leading to a more developed race lets look at some very basic statistics for all of the illiterate 12 year old retards who dont understand how to interpret anything that doesnt tell them UR A GOOD PLAYER U JUST LOST CUS THE GAME IS IMBALANCED number of players at a given level of play does not correlate with balance. AT ALL. lets say fayth made a post in sc2 strategy saying he would give 1000$ to every player who earned 1500 rating as zerg. boy these graphs would look alot different wouldnt they. does that mean that zerg is all of a sudden overpowered and stronger then terran? no. the dumbest thing by far in this thread is the 1600 and 1500 rating level graph, by far. anyone whos done anything in statistics (or has half a brain) knows that a sample size of 19 (YES THERE ARE 19 PLAYERS ABOVE 1500) is not an adequate sample size to judge something as complex as game balance. IM A MORON, WHY IS THIS PETZERGLING? each 1 person on the graph represents over 5% of the total population. lets say a T player with 1501 rating plays a Z player with 1499 rating and the Z player wins (LOL CANT HAPPEN TERRAN OP) this brings an adjustment to our graph of over 10% relative data. considering people at this level play upwards of 10 games a day. proof that this sample size is retarded? here i just re calculated the numbers courtesy of sc2ranks.com and put them on a graph OMFG PROTOSS UPSWING NOW PROTOSS IS IMBALANCED. no the race of players in the top 1500 fluctuates this much on a daily(hourly?) basis k im done see you idiots in disneyland User was temp banned for this post.
Heh, it's pretty clear you're a losing Terran player frustrated because nerfs will make you lose more. Everyone knows there are some balance issues with Terran.
|
On September 02 2010 14:33 WeSt wrote:This graph is ridiculous I mean... there are like 20 players in the world above 1500 points AND out of those 20 there are 9 protosses, 7 terrans, 3 zergs and 1 random. I actually took the time to take a look at the sc2ranks link you posted and it prooves you either made a mistake while making the graphic or you manipulated the information. here I show u picz:
I'm afraid that the much linked Race Distribution Graph which you can see on SC2Ranks is not up to date. That's where you're getting your "lol only 7 terrans" from. If you actually look at the raw data on the main page you will see the following:
10 Protoss above 1500 (3 above 1600)
11 Terran above 1500 (2 above 1600)
4 Zerg above 1500 (1 above 1600)
+1 Random in the form of David Kim also above 1500.
It is however rather easy to drop even 100 points with a few losses, when at these high ranks you lose a lot of points for losing and win very few for winning. It would probably better to amalgamate everyone above 1400 or even everyone above 1300.
|
I don't ever post but all the nonsense here is frustrating me. For one, the top point brackets DO NOT HAVE SAMPLE SIZES. Once again, they are not samples - they are actual data on the entire population. You can't complain about a small sample if the population itself is the exact same size. How is this a difficult concept at all to grasp?
Obviously there wont be a lot of 1600+ players. That's the point.
|
I can not believe how stupid many people are in this thread, claiming the sample size is not big enough.
First, as has been stated many many times, it is no sample, it IS THE ACTUAL population.
Second (i learned this from playing poker): If you are to play a game of flipping coins, and you are allowed to flip the coin once before the game starts, and it is heads, would you rather chose heads for the duration of the game or tails? Given no other information, the obvious answer would be heads, since even though the sample is (very) small in this case, the best prediction you can make is use that information instead of stupidly claiming it is obsolete and it doesnt matter what side you chose.
|
Looks to me like terran definately has a biased. Think about it this way. On every rating and level until the very top the win/loss ratio is 50/50 because the player skill is shifted by the race. A deservingly 1000 rated Terran will have a 1200 rating and face players that are outright better than they are. an 800 will face 1000 level players, 800 will face 600 etc. For most of those levels the population will be representative of the SC2 population as a whole.
This breaks down on the highest level of play where there simply aren't any 1500 rated Zerg/Protoss to face the 1300 level terrans. Therefore at top levels Terrans dominate because there aren't any players who are outright better than they are to mix in and keep a “fair” score.
Its as if the entire population of SC2 is shifted 200 points in favor of Terran.
|
LOL at all the T players at the end. And again, random is obviously OP. :p
|
The graph in the OP surely does look worrying but after seeing the absolute number graph, it doesn't look that bad. Sample size seems really quite small in the upper level.
|
Protoss OP at 600 points obviously
But it would seem that there are just more active "experts" playing terran atm
|
Looks about right to me. The balance isn't bad until you are really good and really understand how to exploit the issues. Balance is always going to be hardest to tweak properly at the highest levels of play.
|
On September 02 2010 23:21 ggrrg wrote:
The graph in the OP surely does look worrying but after seeing the absolute number graph, it doesn't look that bad. Sample size seems really quite small in the upper level.
its not a sample its the population (which was stated 10 times in this thread).
but still... the chart doesnt show anything at all IMO. there are many possible meanings:
1) good players just happen to like playing terran
2) terran is easyer to master than the other races
3) the other races havent yet figured out their matchups (especially vT)
4) terran is overpowered
maybe its a little bit of everything but I think the main reason comes from reason nr3.
|
On September 02 2010 08:08 Doomrok wrote:Terran is perfectly balanced, I mean, look at my Terran vs Zerg win rate! http://www.danrok.com/stats/r r:r
And your ZvT is high aswell, what's your point?
|
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced.
Could this graph by rescaled via (n * percentwithin subgroup)? That way we can see that the "huge terran imbalance" is only in a tiny population that none of us will be part of.
Y'all are freaking monkeys if you didn't notice the fact that the population subsets these are pulled out of are VASTLY DISPARATE.
EDIT: Upon doing calculations, this graph actually shows that Protoss is the most overrepresented race above 600. In fact, it is statistically significantly higher than Terran or Zerg. WHOOPS! Looks like we should be discussing Protoss IMBA?
|
Arghhhhh whether terran is extremely overpowered or not, these stats mean nothing. It's a representation of how many players are playing each race.
Lots of top players being terran =/= terran stronger than the other races.
Obviously the game needs balancing but you can't prove it with numbers. With the way matchmaking works, and the point system, we can not provide any reasonable arguments. We KNOW the game needs balancing because of the games we've watched and the strategies we have seen and the counters to those strategies we've come up with. Blizzard also knows this, which is why they are releasing a balance patch...
Im probably not the only one sick of people trying to make it look worse than it actually is with these retarded numbers.
Does anyone remember in Beta phase 1 when terran was weak? Blizzard openly admitted Terran was doing poorly compared to the other races. Sure the race was buffed throughout the beta test, but what also happened was that players learned how to use all the handy Terran mechanics.
|
On September 03 2010 00:05 Sleight wrote:
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced.
The trend holds over 160 players (the top 3 groups, more if you count the decline in Protoss). 20 players is a decent sample size, contrary to the random assertions otherwise. I will see if I have time tonight to actually do the numbers. You can draw statistics from these, and my intuition says there is a fair chunk of significance to them.
And this deserves reposting:
TT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected.
|
Just to complete this topic a bit.
Taken from a 200.000 Players database / Europe. Showing the Win / Loss Ratio and the Amount of Races within a Pointrange.
Diamond only:
Pointrange: 500 - 600
Terran Players: 399 (Win/Loss Ratio: 55.8325814536)
Protoss Players: 464 (Win / Loss Ratio: 55.6056034483)
Zerg Players: 302 (Win / Loss Ratio: 54.6768211921)
Random Players: 121 (Win / Loss Ratio: 54.6049586777)
Pointrange: 600 - 700
Terran Players: 535 (Win/Loss Ratio: 55.4304672897)
Protoss Players: 678 (Win / Loss Ratio: 55.2837758112)
Zerg Players: 514 (Win / Loss Ratio: 54.9593385214)
Random Players: 164 (Win / Loss Ratio: 54.662195122)
Pointrange: 700 - 800
Terran Players: 412 (Win/Loss Ratio: 55.4825242718)
Protoss Players: 477 (Win / Loss Ratio: 54.8161425577)
Zerg Players: 420 (Win / Loss Ratio: 54.7042857143)
Random Players: 117 (Win / Loss Ratio: 54.3222222222)
Pointrange: 800 - 900
Terran Players: 255 (Win/Loss Ratio: 55.7874509804)
Protoss Players: 363 (Win / Loss Ratio: 54.9438016529)
Zerg Players: 278 (Win / Loss Ratio: 55.1960431655)
Random Players: 66 (Win / Loss Ratio: 54.7045454545)
Pointrange: 900 - 1000
Terran Players: 173 (Win/Loss Ratio: 56.2150289017)
Protoss Players: 208 (Win / Loss Ratio: 55.2403846154)
Zerg Players: 167 (Win / Loss Ratio: 55.0862275449)
Random Players: 31 (Win / Loss Ratio: 54.8774193548)
Pointrange: 1000 - 1100
Terran Players: 98 (Win/Loss Ratio: 57.0183673469)
Protoss Players: 106 (Win / Loss Ratio: 55.408490566)
Zerg Players: 76 (Win / Loss Ratio: 55.6763157895)
Random Players: 11 (Win / Loss Ratio: 55.0272727273)
Pointrange: 1100 - 1200
Terran Players: 51 (Win/Loss Ratio: 57.1196078431)
Protoss Players: 50 (Win / Loss Ratio: 56.056)
Zerg Players: 33 (Win / Loss Ratio: 56.0575757576)
Random Players: 3 (Win / Loss Ratio: 59.4)
Pointrange: 1200 - 9999
Terran Players: 49 (Win/Loss Ratio: 60.2224489796)
Protoss Players: 32 (Win / Loss Ratio: 58.74375)
Zerg Players: 21 (Win / Loss Ratio: 57.4333333333)
Random Players: 2 (Win / Loss Ratio: 59)
And sorry. No Graph. Maybe someone wants to make one 
edit:
now with "Random" Players.
|
On September 03 2010 00:21 kunstprodukt wrote:Just to complete this topic a bit. Taken from a 200.000 Players database / Europe. Showing the Win / Loss Ratio and the Amount of Races within a Pointrange. Diamond only: Pointrange: 500 - 600 Terran Players: 399 (Win/Loss Ratio: 55.8325814536) Protoss Players: 464 (Win / Loss Ratio: 55.6056034483) Zerg Players: 302 (Win / Loss Ratio: 54.6768211921) Pointrange: 600 - 700 Terran Players: 535 (Win/Loss Ratio: 55.4304672897)| Protoss Players: 678 (Win / Loss Ratio: 55.2837758112)| Zerg Players: 514 (Win / Loss Ratio: 54.9593385214) Pointrange: 600 - 700 Terran Players: 535 (Win/Loss Ratio: 55.4304672897)| Protoss Players: 678 (Win / Loss Ratio: 55.2837758112)| Zerg Players: 514 (Win / Loss Ratio: 54.9593385214) Pointrange: 700 - 800 Terran Players: 412 (Win/Loss Ratio: 55.4825242718)| Protoss Players: 477 (Win / Loss Ratio: 54.8161425577)| Zerg Players: 420 (Win / Loss Ratio: 54.7042857143) Pointrange: 800 - 900 Terran Players: 255 (Win/Loss Ratio: 55.7874509804)| Protoss Players: 363 (Win / Loss Ratio: 54.9438016529)| Zerg Players: 278 (Win / Loss Ratio: 55.1960431655) Pointrange: 900 - 1000 Terran Players: 173 (Win/Loss Ratio: 56.2150289017)| Protoss Players: 208 (Win / Loss Ratio: 55.2403846154)| Zerg Players: 167 (Win / Loss Ratio: 55.0862275449) Pointrange: 1000 - 1100 Terran Players: 98 (Win/Loss Ratio: 57.0183673469)| Protoss Players: 106 (Win / Loss Ratio: 55.408490566)| Zerg Players: 76 (Win / Loss Ratio: 55.6763157895) Pointrange: 1100 - 1200 Terran Players: 51 (Win/Loss Ratio: 57.1196078431)| Protoss Players: 50 (Win / Loss Ratio: 56.056)| Zerg Players: 33 (Win / Loss Ratio: 56.0575757576) Pointrange: 1200 - 9999 Terran Players: 49 (Win/Loss Ratio: 60.2224489796)| Protoss Players: 32 (Win / Loss Ratio: 58.74375)| Zerg Players: 21 (Win / Loss Ratio: 57.4333333333) And sorry. No Graph. Maybe someone wants to make one
Im confused what was the exact website where this info came from? i see no random in this. Does this website count random as whatever race they spawned in that specific game?
|
On September 03 2010 00:21 kunstprodukt wrote:Just to complete this topic a bit. Taken from a 200.000 Players database / Europe. Showing the Win / Loss Ratio and the Amount of Races within a Pointrange. Diamond only: Pointrange: 500 - 600 Terran Players: 399 (Win/Loss Ratio: 55.8325814536) Protoss Players: 464 (Win / Loss Ratio: 55.6056034483) Zerg Players: 302 (Win / Loss Ratio: 54.6768211921) Pointrange: 600 - 700 Terran Players: 535 (Win/Loss Ratio: 55.4304672897)| Protoss Players: 678 (Win / Loss Ratio: 55.2837758112)| Zerg Players: 514 (Win / Loss Ratio: 54.9593385214) Pointrange: 600 - 700 Terran Players: 535 (Win/Loss Ratio: 55.4304672897)| Protoss Players: 678 (Win / Loss Ratio: 55.2837758112)| Zerg Players: 514 (Win / Loss Ratio: 54.9593385214) Pointrange: 700 - 800 Terran Players: 412 (Win/Loss Ratio: 55.4825242718)| Protoss Players: 477 (Win / Loss Ratio: 54.8161425577)| Zerg Players: 420 (Win / Loss Ratio: 54.7042857143) Pointrange: 800 - 900 Terran Players: 255 (Win/Loss Ratio: 55.7874509804)| Protoss Players: 363 (Win / Loss Ratio: 54.9438016529)| Zerg Players: 278 (Win / Loss Ratio: 55.1960431655) Pointrange: 900 - 1000 Terran Players: 173 (Win/Loss Ratio: 56.2150289017)| Protoss Players: 208 (Win / Loss Ratio: 55.2403846154)| Zerg Players: 167 (Win / Loss Ratio: 55.0862275449) Pointrange: 1000 - 1100 Terran Players: 98 (Win/Loss Ratio: 57.0183673469)| Protoss Players: 106 (Win / Loss Ratio: 55.408490566)| Zerg Players: 76 (Win / Loss Ratio: 55.6763157895) Pointrange: 1100 - 1200 Terran Players: 51 (Win/Loss Ratio: 57.1196078431)| Protoss Players: 50 (Win / Loss Ratio: 56.056)| Zerg Players: 33 (Win / Loss Ratio: 56.0575757576) Pointrange: 1200 - 9999 Terran Players: 49 (Win/Loss Ratio: 60.2224489796)| Protoss Players: 32 (Win / Loss Ratio: 58.74375)| Zerg Players: 21 (Win / Loss Ratio: 57.4333333333) And sorry. No Graph. Maybe someone wants to make one
I love you for this Data! Hope this shows everyone how tiny the imbalance is atm. I accept that Terran needs some small Tweaks, but most of the time you should be better looking at yourself then on the balance
|
Terran just has more options, which means that higher level players will be able to exploit them much more efficiently than lower level players.
Thats my hypothesis of why there are more Terran wins at higher levels.
I think Zerg, Protoss should be buffed by giving them more options, possibly new units but realistically new abilities. It would not affect lower levels too much due to the APM requirement and would help tremendously in the higher levels.
Something as simple as giving Warp Prism Energy Fields a boost to attack speed for Protoss units. Or adding an upgrade to make Transfuse AoE and also give a big speed increase for Queens. How often do you see Queens used on the offensive?
|
On September 02 2010 23:07 Duban wrote:
Looks to me like terran definately has a biased. Think about it this way. On every rating and level until the very top the win/loss ratio is 50/50 because the player skill is shifted by the race. A deservingly 1000 rated Terran will have a 1200 rating and face players that are outright better than they are. an 800 will face 1000 level players, 800 will face 600 etc. For most of those levels the population will be representative of the SC2 population as a whole.
This breaks down on the highest level of play where there simply aren't any 1500 rated Zerg/Protoss to face the 1300 level terrans. Therefore at top levels Terrans dominate because there aren't any players who are outright better than they are to mix in and keep a “fair” score.
Its as if the entire population of SC2 is shifted 200 points in favor of Terran.
Then terran should also get more points than other races, cause they fight stronger enemies. While losing less points and winning more AND having the same win ratio as other race, they should have 200 average points more.
RIght now (global Diamond):
Protoss 581
Terran 581
Zerg 585
But i guess someone will manage to argue why terran is op because of zerg having the most points.
On September 03 2010 00:28 Zombo Joe wrote:
Terran just has more options, which means that higher level players will be able to exploit them much more efficiently than lower level players.
Thats my hypothesis of why there are more Terran wins at higher levels.
Sounds reasonably.
|
On September 03 2010 00:25 tipakee wrote:
Im confused what was the exact website where this info came from? i see no random in this. Does this website count random as whatever race they spawned in that specific game?
The Data is not published, because it was crawled by my own bots.
Well, Random is within the DB, but i was not sure if it would be interesting to have it in the table here. So i skipped it.
|
On September 03 2010 00:29 kunstprodukt wrote:Show nested quote +On September 03 2010 00:25 tipakee wrote:
Im confused what was the exact website where this info came from? i see no random in this. Does this website count random as whatever race they spawned in that specific game?
The Data is not published, because it was crawled by my own bots. Well, Random is within the DB, but i was not sure if it would be interesting to have it in the table here. So i skipped it.
O ok, just checking since i'm a random player myself and the numbers just intrest me. thanks for the time to composite your table
|
Agreed with Tyler... The end of the graph is not a sample, it's the entire population. How many people actually studied statistics past high school here?
What this graph means is "there is more terran top players than any other race, zerg is the worst". You draw your conclusions from here. At the very worst cut 1% at each end and you'll still get very smilar stats.
My conclusion is: the higher you are in diamond league, the higher the difference "little" imabalances make, and favor terran.
And I play random.
People at 1500 diamond don't play terran because "it's the campaign race" omg.
|
@kunstprodukt if you present this data as scv I would like to see it, I thought about doing a little crawling myself, maybe we can work together. Do you want to do a little query at your database and look at the matchups, I have not done so because I am not suceeding in getting enough profiles (also it increses the possibility of spam prevention ban a lot).
*Edit: I would like to see the data as CSV, not as SCV though I would look at a SCV with the data written on it.*
|
On September 03 2010 00:31 tipakee wrote:Show nested quote +On September 03 2010 00:29 kunstprodukt wrote:On September 03 2010 00:25 tipakee wrote:
Im confused what was the exact website where this info came from? i see no random in this. Does this website count random as whatever race they spawned in that specific game?
The Data is not published, because it was crawled by my own bots. Well, Random is within the DB, but i was not sure if it would be interesting to have it in the table here. So i skipped it. O ok, just checking since i'm a random player myself and the numbers just intrest me. thanks for the time to composite your table
did an "edit" and added random players.
|
On September 03 2010 00:38 ReplayArk wrote:
@kunstprodukt if you present this data as scv I would like to see it, I thought about doing a little crawling myself, maybe we can work together. Do you want to do a little query at your database and look at the matchups, I have not done so because I am not suceeding in getting enough profiles (also it increses the possibility of spam prevention ban a lot).
*Edit: I would like to see the data as CSV, not as SCV though I would look at a SCV with the data written on it.*
Sure.
Just let me know in which format you need the datas.. I will format them like that.
|
If you run the stats on Sc2ranks.com it shows 977720 players and 6.72% are in diamond. There are 188% more terrans than Zerg in the entire population, but only 128% more in diamond. Based on the entire population there are 123% more Zerg in diamond than there should be based on the number of Zerg playing and that 6.72% of all players are in diamond. Terrans have only 84% in diamond. Protoss are actually pretty close at 98%. I am not suggesting Zerg are op, or that Terran are up. I posted this to point out that numbers taking only a single factor into account don't give you a big picture. The graph is interesting and is something to be contemplated, but once you start adding other factors you get different results. With constantly changing strategies you can't take a snapshot of data and say it is how it always is.
|
On September 03 2010 00:42 BlackIce81 wrote:
If you run the stats on Sc2ranks.com it shows 977720 players and 6.72% are in diamond. There are 188% more terrans than Zerg in the entire population, but only 128% more in diamond. Based on the entire population there are 123% more Zerg in diamond than there should be based on the number of Zerg playing and that 6.72% of all players are in diamond. Terrans have only 84% in diamond. Protoss are actually pretty close at 98%. I am not suggesting Zerg are op, or that Terran are up. I posted this to point out that numbers taking only a single factor into account don't give you a big picture. The graph is interesting and is something to be contemplated, but once you start adding other factors you get different results. With constantly changing strategies you can't take a snapshot of data and say it is how it always is.
You do realize Plat and below doesn't even have the basics down right?
Stuff like "build a supply depot before you're supply capped" is stuff they're still trying to get the concept of.
|
Some of the first posters are correct. The data of the top-top diamonds is statistically insignificant because of the very small number of players at this level.
Basically the graph becomes less representative of reality the higher you go up the point score. It 'could' be accurate, but this data does not prove that, so the people QQing about Terran so OP should probably learn a little statistics. Doesn't matter that it's the "entire population." If that's your reasoning you have no clue what you're talking about here.
|
If twice as many people play Terran we can expect there to be more terrans than Zerg in diamond.
|
On September 03 2010 01:06 Tray wrote:
Some of the first posters are correct. The data of the top-top diamonds is statistically insignificant because of the very small number of players at this level.
Basically the graph becomes less representative of reality the higher you go up the point score. It 'could' be accurate, but this data does not prove that, so the people QQing about Terran so OP should probably learn a little statistics. Doesn't matter that it's the "entire population." If that's your reasoning you have no clue what you're talking about here.
There will always be a very small number of players at high levels. You're as good as saying that it's impossible to use statistics to measure anything significant, since the match-maker will by definition even out populations at more casual levels.
|
On September 03 2010 01:40 Karkadinn wrote:Show nested quote +On September 03 2010 01:06 Tray wrote:
Some of the first posters are correct. The data of the top-top diamonds is statistically insignificant because of the very small number of players at this level.
Basically the graph becomes less representative of reality the higher you go up the point score. It 'could' be accurate, but this data does not prove that, so the people QQing about Terran so OP should probably learn a little statistics. Doesn't matter that it's the "entire population." If that's your reasoning you have no clue what you're talking about here. There will always be a very small number of players at high levels. You're as good as saying that it's impossible to use statistics to measure anything significant, since the match-maker will by definition even out populations at more casual levels.
At the top level, with small playerbase, yes it is impossible. That's exactly the point. You need to have a decent number of players before you're getting a clear view of the actual situation.
For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis.
|
On September 03 2010 00:28 Zombo Joe wrote:
Terran just has more options, which means that higher level players will be able to exploit them much more efficiently than lower level players.
Thats my hypothesis of why there are more Terran wins at higher levels.
I think Zerg, Protoss should be buffed by giving them more options, possibly new units but realistically new abilities. It would not affect lower levels too much due to the APM requirement and would help tremendously in the higher levels.
Something as simple as giving Warp Prism Energy Fields a boost to attack speed for Protoss units. Or adding an upgrade to make Transfuse AoE and also give a big speed increase for Queens. How often do you see Queens used on the offensive?
Queens were used offensively in beta at one point. They were nerfed as a result. Apparently Blizzard didn't like Zerg having more than one offensive caster unit. I don't see Blizzard going back on that, either... it's surpassingly rare for them to ever roll back changes once they're made, even if the reasons for those changes have stopped existing. Maybe in an expansion, but definitely not in a regular patch.
|
United States59 Posts
Racial Distributions at the Diamond level: Nerdy Asians and Whites lolol
|
On September 03 2010 01:44 Tray wrote:Show nested quote +On September 03 2010 01:40 Karkadinn wrote:On September 03 2010 01:06 Tray wrote:
Some of the first posters are correct. The data of the top-top diamonds is statistically insignificant because of the very small number of players at this level.
Basically the graph becomes less representative of reality the higher you go up the point score. It 'could' be accurate, but this data does not prove that, so the people QQing about Terran so OP should probably learn a little statistics. Doesn't matter that it's the "entire population." If that's your reasoning you have no clue what you're talking about here. There will always be a very small number of players at high levels. You're as good as saying that it's impossible to use statistics to measure anything significant, since the match-maker will by definition even out populations at more casual levels. At the top level, with small playerbase, yes it is impossible. That's exactly the point. You need to have a decent number of players before you're getting a clear view of the actual situation. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis.
You're taking the absolute extreme to support your point. Even if you just stop at 1200, the differences are sufficiently obvious. Trends among the top players are also important because one faction mirror matches do not make particularly good media entertainment for tournaments.
|
On September 02 2010 07:55 Liquid`Tyler wrote:
As far as I know, sc2ranks is pretty damn comprehensive of Diamond, especially high Diamond (where it seems some people have a problem with the "sample size"). Yeah, the number of people in the 1500+ group is small but that doesn't mean that there is a problem with the sample size. These numbers aren't extrapolated from a small population of the 1500+ Diamond group. These numbers directly represent that group.
Warning: This post is irrefutable.
This graph is terribly misleading and is a great example of how statistics can be used improperly to support an agenda.
Total Sample size = 27862
number of people in 1200-1300 range = 180
number of people in 1300-1400 range = 101
number of people in 1400-1500 range = 38
number of people in 1500+ range = 20
Sum of 1200+ = 339
Percentage of 600+ diamond players that are 1200+
= 339/27862 = 1.2%
Displaying a total sample size of 28k then attempting to draw trending conclusions from a trend that only appears once you've reached outliers that make up 1.2% of the sample size is just terrible statistics work. If this was done in industry you'd lose your job in a day and potentially face a lawsuit. You just DON'T draw conclusions based on statistical outliers. Thats like saying it snows a lot in Texas based on one inch of snowfall every 3 years.
|
I think a lot of people in this thread are missing a logical connection. No one is really trying to say that Terran is imba because more people picked them at a high level. People are using that graph as further evidence that Terran might be imba because it indicates not that Terran is more chosen by high level players, but that you're more likely to be a high rate player if you choose Terran (as a product of everyone being around a 50% win rate at the level you're supposed to be at).
|
I enjoy the references by people of the "sample size being too small", when the sample size is equal to 100% of the population.
For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis.
Everyone was previously saying we should only be balancing at the top of the game, as if you're not near the top, you can advance by getting better. Looking at 1200+ Diamond, a clear pattern is emerging. Terran's are dominating the top 400 players.
That being said, this graph proves what has already been said multiple times, Terrans have an advantage at the top and Zerg is weak. That's why Blizzard is releasing Patch 1.1
|
On September 03 2010 01:38 BlackIce81 wrote:
If you run the stats on Sc2ranks.com it shows 977720 players and 6.72% are in diamond. There are 188% more terrans than Zerg in the entire population, but only 128% more in diamond. Based on the entire population there are 123% more Zerg in diamond than there should be based on the number of Zerg playing and that 6.72% of all players are in diamond. Terrans have only 84% in diamond. Protoss are actually pretty close at 98%. I am not suggesting Zerg are op, or that Terran are up. I posted this to point out that numbers taking only a single factor into account don't give you a big picture. The graph is interesting and is something to be contemplated, but once you start adding other factors you get different results. With constantly changing strategies you can't take a snapshot of data and say it is how it always is.
How about this, roughly 30% of diamond players are terran (according to sc2ranks.com). In the NA, EU and KR top 200 rankings blizzard releases terran are over 40%, meanwhile protoss and zerg both are underrepresented in top 200 compared to their presence in diamond league.
Good players will reach diamond no matter what race they play. Looking at race %-ages in lower leagues can be iffy, especially since terran is MASSIVELY overrepresented in bronze league (33.32% in silver, 43.59% in bronze), likely due to campaign being terran which makes most new players play terran, you couldn't play zerg at all in campaign which is likely why zerg is slightly underrepresented in bronze league. Bronze also has the highest numbers of players by far which further messes up your %-ages, something like 40% of the total terran population is in bronze.
|
Your rating is not determined by how skilled you are directly. It is determined by wins and losses exclusively.
A 1600 rated diamond zerg player might be more skilled than a 1600 rated diamond terran player. They are only the same rating because they win a near equal amount.
You can't use ratings to judge skill in this way.
There are two problems with trying to analyze things statistically in this game:
A) 99.9% of players lack the skill to be used as an example of balance.
B) .1% of players are too small a number to tell if it's a fluke or if it's legitimate.
Btw, when people say "sample size" they are actually correct. 20 players are not a good sample size of the game as a whole. Even though those 20 players are 100% of the 1500+ group, they are only a VERY SMALL fraction of the total population of the game. So their size is not a good sample of the entire game's population.
You might say "that doesn't matter, because they are the skilled ones". Yes, they are the skilled ones, but their race choice can't be seen as something significant. For example, Idra says Zerg is underpowered vs. Terran. Yet he still plays Zerg. That throws the entire data set out the window.
That half of players who are Terran at 1500+ might play Terran JUST BECAUSE THEY PREFER THAT RACE, and the same probably goes for the rest of the players there.
Therefor, balance can't be determined by population. It makes no sense.
|
Just because the data is exactly what you would expect to see if there was an imbalance...
Just because the game developer has come out and admitted there IS an imbalance...
Just because the game developer has already scheduled a patch to address the imbalance...
DOES NOT MEAN THERE IS AN IMBALANCE!!!
Terran are just better players. Zerg doesn't get it, Protoss doesn't get it, Pros don't get it, Blizzard doesn't get it...
They are all scrubs who need to L2P and there is ZERO evidence that would sway my opinion.
If anything Terran is underpowered and the patch is a liberal conspiracy by the Obama administration.
I will now reinsert my ear plugs, blinders and go back to my fantasy world where all things are shiny and happy and SC2 is perfectly balanced.
|
27k is a sample size. The actual data contains close to a million records. This graph ignores 93% of the data. When twice as many people play Terran you are going to end up with more terrans in diamond. I am not saying terrans don't need nerfed. I am saying this graph does not directly support the fact that terrans are strong right now. If you use the same data but break it down by 50 points instead of 100, the top three has one of each race.
|
something like 40% of the total terran population is in bronze.
34% total Zerg are bronze as well. 8% are diamond compared to terrans which only have 5.5%.
|
these graphs are assuming that there are an equal number of zergs, terrans, and protoss's in diamond league to begin with.
|
On September 03 2010 02:11 Buddhist wrote:
Therefor, balance can't be determined by population. It makes no sense.
Relative population combined with trends gives a pretty good idea though. As I pointed out terran is overrepresented in the top 200 compared to population, in addition terran is also trending upwards while both zerg and protoss are trending downwards in top 200 lists.
Personal preferences and skill will always skew the statistics one way or the other though.
|
On September 03 2010 02:11 Buddhist wrote:
There are two problems with trying to analyze things statistically in this game:
A) 99.9% of players lack the skill to be used as an example of balance.
B) .1% of players are too small a number to tell if it's a fluke or if it's legitimate.
Your first point is wrong because imbalance plays a role even in matches between less skilled players. Your second is wrong because, among other reasons, you don't have to look at .1% of the chart to see the trends.
|
Sanya12364 Posts
Percentages aren't all that helpful when there is uneven numbers at all levels of diamond.
You want to see where there is mass. Histograms would serve a better purpose. Log scaling on the side would also be necessary since the buckets vary so much in magnitude.
BTW. This is much improved over previous analysis (lol @ only five skill buckets). Are we really suppose to think that all platinum players are all the same or that there is no overlap between leagues.
It still has some the flaws of any sc2ranks analysis in that you only know a player's favorite race (most played race.)
|
On September 03 2010 02:25 biology]major wrote:
these graphs are assuming that there are an equal number of zergs, terrans, and protoss's in diamond league to begin with.
These graphs are assuming that players who win the most are at the top.
Strangely enough they are...
Why are they winning more?
That is the question that has already been answered and is currently being addressed by the developers.
|
On September 03 2010 01:53 Karkadinn wrote:Show nested quote +On September 03 2010 01:44 Tray wrote:On September 03 2010 01:40 Karkadinn wrote:On September 03 2010 01:06 Tray wrote:
Some of the first posters are correct. The data of the top-top diamonds is statistically insignificant because of the very small number of players at this level.
Basically the graph becomes less representative of reality the higher you go up the point score. It 'could' be accurate, but this data does not prove that, so the people QQing about Terran so OP should probably learn a little statistics. Doesn't matter that it's the "entire population." If that's your reasoning you have no clue what you're talking about here. There will always be a very small number of players at high levels. You're as good as saying that it's impossible to use statistics to measure anything significant, since the match-maker will by definition even out populations at more casual levels. At the top level, with small playerbase, yes it is impossible. That's exactly the point. You need to have a decent number of players before you're getting a clear view of the actual situation. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. You're taking the absolute extreme to support your point. Even if you just stop at 1200, the differences are sufficiently obvious. Trends among the top players are also important because one faction mirror matches do not make particularly good media entertainment for tournaments.
Even in the 1200+ the stats are still probably not statistically relevant. As someone posted, that's less than 1.5% of all diamond players. I'm not gonna do the math, just trying to enlighten some people on stats here because there's a lot of sheep jumping to horribly retarded conclusions.
|
On September 03 2010 01:58 Fraud wrote:I enjoy the references by people of the "sample size being too small", when the sample size is equal to 100% of the population. Show nested quote +For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. Everyone was previously saying we should only be balancing at the top of the game, as if you're not near the top, you can advance by getting better. Looking at 1200+ Diamond, a clear pattern is emerging. Terran's are dominating the top 400 players. That being said, this graph proves what has already been said multiple times, Terrans have an advantage at the top and Zerg is weak. That's why Blizzard is releasing Patch 1.1
This is not true and the person who posted this is not very smart. Please don't post on statistics if you don't understand it.
|
On September 02 2010 07:57 Mikilatov wrote:
Pretty eye-opening, it seems.
I'm glad that this graph pulls up an interesting point though, Terrans aren't really that overpowered except at high levels in the hands of 1000+ point diamond players.
I am sorry but you are a moron. All being overpowered does is say ok if your terran and you play enough games you are going to be 100 points higher rated then if you're protoss and another 100 points higher then if your zerg. Of course the percentages will even out Terrans are just higher ranked then their skill would otherwise allow them to be. Terrans will have roughly the same win percentages too because as a result of winnings games they couldnt with zerg they now have to play harder opponents.
|
On September 03 2010 02:51 Tray wrote:Show nested quote +On September 03 2010 01:53 Karkadinn wrote:On September 03 2010 01:44 Tray wrote:On September 03 2010 01:40 Karkadinn wrote:On September 03 2010 01:06 Tray wrote:
Some of the first posters are correct. The data of the top-top diamonds is statistically insignificant because of the very small number of players at this level.
Basically the graph becomes less representative of reality the higher you go up the point score. It 'could' be accurate, but this data does not prove that, so the people QQing about Terran so OP should probably learn a little statistics. Doesn't matter that it's the "entire population." If that's your reasoning you have no clue what you're talking about here. There will always be a very small number of players at high levels. You're as good as saying that it's impossible to use statistics to measure anything significant, since the match-maker will by definition even out populations at more casual levels. At the top level, with small playerbase, yes it is impossible. That's exactly the point. You need to have a decent number of players before you're getting a clear view of the actual situation. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. You're taking the absolute extreme to support your point. Even if you just stop at 1200, the differences are sufficiently obvious. Trends among the top players are also important because one faction mirror matches do not make particularly good media entertainment for tournaments. Even in the 1200+ the stats are still probably not statistically relevant. As someone posted, that's less than 1.5% of all diamond players. I'm not gonna do the math, just trying to enlighten some people on stats here because there's a lot of sheep jumping to horribly retarded conclusions.
Good luck convincing all the 'sheep' of your pov when your methodology consists of declaring all information beyond an arbitrary self-determined point completely meaningless.
|
Perfect representation of how race representation represents the overall race population at the lower levels. As you get toward the higher levels where it is no longer a matter of better Z/P hanging with inferior T, the data skews HEAVILY in favor of T.
Thanks for the data. Statistically speaking we can now say "with confidence " that Terran is out of whack.
|
So much denial. Yeah, there might only be a few guys in the top 0.1%, making the last bars more inaccurate than the earlier ones.
But, the TREND of the data speaks volumes.
|
On September 03 2010 02:51 Tray wrote:Show nested quote +On September 03 2010 01:53 Karkadinn wrote:On September 03 2010 01:44 Tray wrote:On September 03 2010 01:40 Karkadinn wrote:On September 03 2010 01:06 Tray wrote:
Some of the first posters are correct. The data of the top-top diamonds is statistically insignificant because of the very small number of players at this level.
Basically the graph becomes less representative of reality the higher you go up the point score. It 'could' be accurate, but this data does not prove that, so the people QQing about Terran so OP should probably learn a little statistics. Doesn't matter that it's the "entire population." If that's your reasoning you have no clue what you're talking about here. There will always be a very small number of players at high levels. You're as good as saying that it's impossible to use statistics to measure anything significant, since the match-maker will by definition even out populations at more casual levels. At the top level, with small playerbase, yes it is impossible. That's exactly the point. You need to have a decent number of players before you're getting a clear view of the actual situation. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. You're taking the absolute extreme to support your point. Even if you just stop at 1200, the differences are sufficiently obvious. Trends among the top players are also important because one faction mirror matches do not make particularly good media entertainment for tournaments. Even in the 1200+ the stats are still probably not statistically relevant. As someone posted, that's less than 1.5% of all diamond players. I'm not gonna do the math, just trying to enlighten some people on stats here because there's a lot of sheep jumping to horribly retarded conclusions.
Dude. This isn't a survey about what color you like. Top players KNOW WHAT'S GOING ON, whereas lower level player DO NOT (as much). You're trying to say that 98.5% of people think that the moon is made of cheese and that 1.5% of people who actually know what the hell they are talking about are irrelevant. This isn't statistics, it's logic.
|
On September 03 2010 00:18 Sentient wrote:Show nested quote +On September 03 2010 00:05 Sleight wrote:
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced. The trend holds over 160 players (the top 3 groups, more if you count the decline in Protoss). 20 players is a decent sample size, contrary to the random assertions otherwise. I will see if I have time tonight to actually do the numbers. You can draw statistics from these, and my intuition says there is a fair chunk of significance to them.And this deserves reposting: Show nested quote +TT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected.
Here is an example of a poster with only the most basic understanding of Statistics. I will just show how this is both wrong and embarrassing.
To this first bolded bit:
Twenty people cannot be significant in a population of 28K. In fact, as Buddhist (i believe) pointed out, the top 3 tiers amount to 160 people, which still remains just over 1 Percent of the population. 1% sample size is not a useful sample size in this kind of sampling analysis unless it is a truly random assortment. The term we are referring to by "sample size" is the "power" of said sample. When you have a small sample, you have less power unless the different you are looking for is HUGE. Actually being involved in professional population anaylses as part of my medical degree, I can say that this sample size LACKS SUFFICIENT POWER to obtain any result given the overall population.
Furthermore, by simply picking the TOP you are eliminating a random element. This is stratified sampling and is inclined to a number of kinds of bias and is not viable unless you can demonstrate that the method is necessary to create a balanced population, This is not the case and is therefore not an acceptable sampling method.
To the second boldded bit:
Based on the graphs presented, of the population of diamond players > 600, Protoss is actually the most prevalent. In fact, a MAJORITY do NOT play Terran. A PLURALITY may play Terran, though not in this sampling, meaning the largest group less than 50%, but over 50% of players, which is the definition of majority, do not play Terran. Not in this graph, not anywhere. Just count them. So to say the number of players >1300 play Terran means nothing unless this is a sampling with adequate power and sample size to demonstrate statistical significance. I have already explained to you that it does not.
Furthermore, you cannot assume TRENDS from a CROSS-SECTIONAL analysis. This type of study only can look at "prevalence" or the actual state of people at this moment. It cannot tell if people are moving up, down, or dying, for all it matters. You can only say "In the 20 person sample size at the highest tier, there are 12 Terran." Or, "in the top 3 tiers of 120 people, Terran comprises a larger than expected statistically significant portion by chi-squared analysis, WITHIN THIS POPULATION." If we try to compare the results to larger populations, we find that they lack statistical significance.
Conclusion:
Sorry, Sentient, one course in community college does not qualify you to be a statistical analyst. And you are wrong. A double whammy. Stop being thick and study harder.
|
On September 03 2010 02:53 Tray wrote:Show nested quote +On September 03 2010 01:58 Fraud wrote:I enjoy the references by people of the "sample size being too small", when the sample size is equal to 100% of the population. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. Everyone was previously saying we should only be balancing at the top of the game, as if you're not near the top, you can advance by getting better. Looking at 1200+ Diamond, a clear pattern is emerging. Terran's are dominating the top 400 players. That being said, this graph proves what has already been said multiple times, Terrans have an advantage at the top and Zerg is weak. That's why Blizzard is releasing Patch 1.1 This is not true and the person who posted this is not very smart. Please don't post on statistics if you don't understand it.
I wrote it before, but I guess I have to write it again:
On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200.
Since you obviously understand statistics, you can check for yourself and provide your numbers. And of course you are allowed to normalize the null hypothesis to account for all these players who just play terran for the looks or because they played them in the campaign. Just tell us the corrected null hypothesis numbers.
|
Yeah, terran seems a bit over-represented.
|
On September 03 2010 03:08 Winter_mute wrote:Show nested quote +On September 03 2010 02:53 Tray wrote:On September 03 2010 01:58 Fraud wrote:I enjoy the references by people of the "sample size being too small", when the sample size is equal to 100% of the population. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. Everyone was previously saying we should only be balancing at the top of the game, as if you're not near the top, you can advance by getting better. Looking at 1200+ Diamond, a clear pattern is emerging. Terran's are dominating the top 400 players. That being said, this graph proves what has already been said multiple times, Terrans have an advantage at the top and Zerg is weak. That's why Blizzard is releasing Patch 1.1 This is not true and the person who posted this is not very smart. Please don't post on statistics if you don't understand it. I wrote it before, but I guess I have to write it again: Show nested quote +On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200. Since you obviously understand statistics, you can check for yourself and provide your numbers. And of course you are allowed to normalize the null hypothesis to account for all these players who just play terran for the looks or because they played them in the campaign. Just tell us the corrected null hypothesis numbers.
You people are so retarded. You're not even accounting for the number of people that play terran relative to the other races. Do you really think your stats are even close to relevant? Wow. Go back to school.
|
On September 03 2010 03:08 Winter_mute wrote:Show nested quote +On September 03 2010 02:53 Tray wrote:On September 03 2010 01:58 Fraud wrote:I enjoy the references by people of the "sample size being too small", when the sample size is equal to 100% of the population. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. Everyone was previously saying we should only be balancing at the top of the game, as if you're not near the top, you can advance by getting better. Looking at 1200+ Diamond, a clear pattern is emerging. Terran's are dominating the top 400 players. That being said, this graph proves what has already been said multiple times, Terrans have an advantage at the top and Zerg is weak. That's why Blizzard is releasing Patch 1.1 This is not true and the person who posted this is not very smart. Please don't post on statistics if you don't understand it. I wrote it before, but I guess I have to write it again: Show nested quote +On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200. Since you obviously understand statistics, you can check for yourself and provide your numbers. And of course you are allowed to normalize the null hypothesis to account for all these players who just play terran for the looks or because they played them in the campaign. Just tell us the corrected null hypothesis numbers.
No one is disagreeing that Terran represents a greater than expected portion of the >1300 population. The argument runs that THIS SAMPLE lacks enough POWER to have STATISTICAL SIGNIFICANCE to the rest of the WHOLE POPULATION.
I capitalized the important words. The null hypothesis here should be "Terran is equally represented as within the greater population" and the same for the other 2 races. We KNOW the greater population, so we can compare BOTH populations directly and see that this population lacks statistical significance when compared to the greater population.
Compute confidence intervals and you will see that there is overlap. If you wouldn't get published, your data is not conclusive.
|
On September 03 2010 02:56 Karkadinn wrote:Show nested quote +On September 03 2010 02:51 Tray wrote:On September 03 2010 01:53 Karkadinn wrote:On September 03 2010 01:44 Tray wrote:On September 03 2010 01:40 Karkadinn wrote:On September 03 2010 01:06 Tray wrote:
Some of the first posters are correct. The data of the top-top diamonds is statistically insignificant because of the very small number of players at this level.
Basically the graph becomes less representative of reality the higher you go up the point score. It 'could' be accurate, but this data does not prove that, so the people QQing about Terran so OP should probably learn a little statistics. Doesn't matter that it's the "entire population." If that's your reasoning you have no clue what you're talking about here. There will always be a very small number of players at high levels. You're as good as saying that it's impossible to use statistics to measure anything significant, since the match-maker will by definition even out populations at more casual levels. At the top level, with small playerbase, yes it is impossible. That's exactly the point. You need to have a decent number of players before you're getting a clear view of the actual situation. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. You're taking the absolute extreme to support your point. Even if you just stop at 1200, the differences are sufficiently obvious. Trends among the top players are also important because one faction mirror matches do not make particularly good media entertainment for tournaments. Even in the 1200+ the stats are still probably not statistically relevant. As someone posted, that's less than 1.5% of all diamond players. I'm not gonna do the math, just trying to enlighten some people on stats here because there's a lot of sheep jumping to horribly retarded conclusions. Good luck convincing all the 'sheep' of your pov when your methodology consists of declaring all information beyond an arbitrary self-determined point completely meaningless.
Who do I have to convince? The burden of proof is on you to prove to me that Terran is imbalanced because of this. Statistically, that has not been shown.
1200+ is just as arbitrary as any other cutoff. You really think diamonds 800-1200 should be ignored? I think Idra falls into that category...
Seriously you people are too dumb to even discuss this topic with at an adult level. I'm guessing none of you have even taken a stats class, let alone graduated high school.
|
On September 03 2010 03:04 whateversclever wrote:Show nested quote +On September 03 2010 02:51 Tray wrote:On September 03 2010 01:53 Karkadinn wrote:On September 03 2010 01:44 Tray wrote:On September 03 2010 01:40 Karkadinn wrote:On September 03 2010 01:06 Tray wrote:
Some of the first posters are correct. The data of the top-top diamonds is statistically insignificant because of the very small number of players at this level.
Basically the graph becomes less representative of reality the higher you go up the point score. It 'could' be accurate, but this data does not prove that, so the people QQing about Terran so OP should probably learn a little statistics. Doesn't matter that it's the "entire population." If that's your reasoning you have no clue what you're talking about here. There will always be a very small number of players at high levels. You're as good as saying that it's impossible to use statistics to measure anything significant, since the match-maker will by definition even out populations at more casual levels. At the top level, with small playerbase, yes it is impossible. That's exactly the point. You need to have a decent number of players before you're getting a clear view of the actual situation. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. You're taking the absolute extreme to support your point. Even if you just stop at 1200, the differences are sufficiently obvious. Trends among the top players are also important because one faction mirror matches do not make particularly good media entertainment for tournaments. Even in the 1200+ the stats are still probably not statistically relevant. As someone posted, that's less than 1.5% of all diamond players. I'm not gonna do the math, just trying to enlighten some people on stats here because there's a lot of sheep jumping to horribly retarded conclusions. Dude. This isn't a survey about what color you like. Top players KNOW WHAT'S GOING ON, whereas lower level player DO NOT (as much). You're trying to say that 98.5% of people think that the moon is made of cheese and that 1.5% of people who actually know what the hell they are talking about are irrelevant. This isn't statistics, it's logic.
Hey champ, it's 1.2% of DIAMOND players. Not 1.2% of all bnet players. That's why it's not significant. Learn to read before you post. I'm diamond 800 with a 70% win rate. By your logic my stats are irrelevant to the balance discussion. I hope you realize the faults in your logic, but you probably don't.
|
I imagine when a fanatical cult sect comes charging into his hometown in order to lynch all of the non-whites in the place, Tray will be standing there, arms crossed, telling them the burden of proof is on them to show that all the non-whites need to be lynched.
They're going to believe what they want to believe until you give them a good reason to believe otherwise. If you genuinely don't care what they believe, then why are you just shouting and screaming at people that they're wrong? If you do care, why are you too lazy to do the math?
That's just textbook trolling right there.
|
On September 03 2010 03:07 Sleight wrote:Show nested quote +On September 03 2010 00:18 Sentient wrote:On September 03 2010 00:05 Sleight wrote:
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced. The trend holds over 160 players (the top 3 groups, more if you count the decline in Protoss). 20 players is a decent sample size, contrary to the random assertions otherwise. I will see if I have time tonight to actually do the numbers. You can draw statistics from these, and my intuition says there is a fair chunk of significance to them.And this deserves reposting: TT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected. Here is an example of a poster with only the most basic understanding of Statistics. I will just show how this is both wrong and embarrassing. To this first bolded bit: Twenty people cannot be significant in a population of 28K. In fact, as Buddhist (i believe) pointed out, the top 3 tiers amount to 160 people, which still remains just over 1 Percent of the population. 1% sample size is not a useful sample size in this kind of sampling analysis unless it is a truly random assortment. The term we are referring to by "sample size" is the "power" of said sample. When you have a small sample, you have less power unless the different you are looking for is HUGE. Actually being involved in professional population anaylses as part of my medical degree, I can say that this sample size LACKS SUFFICIENT POWER to obtain any result given the overall population.
Since you deal in 'professional population analyses' in medicine, just out of interest, how many drugs passed by the FSA or indeed any other international body have been trialed on more than 1% of the population?
Your conclusion is also wrong. You need sample variances to determine power. Sample size is insufficient.
Notwithstanding that, nobody is actually claiming that the 'top X' is representative of the whole population. All the claims refer to balance within the 'top X'. In which case, the sample size is the whole population. Still, to make any claims about 'power' you need sample variances.
|
On September 03 2010 03:20 Bibdy wrote:
I imagine when a fanatical cult sect comes charging into his hometown in order to lynch all of the non-whites in the place, Tray will be standing there, arms crossed, telling them the burden of proof is on them to show that all the non-whites need to be lynched.
They're going to believe what they want to believe until you give them a good reason to believe otherwise. If you genuinely don't care what they believe, then why are you just shouting and screaming at people that they're wrong? If you do care, why are you too lazy to do the math?
That's just textbook trolling right there.
No what I'm doing is telling people why they're wrong. What YOU'RE doing is trolling. You're great at it. You've been doing it since the beta in the bnet forums. I have a job, so I don't have all day to crunch numbers in excel to prove to you that your sample is statistically irrelevant. Go do a little google search on stats, confidence intervals, and the like. You might learn something.
Cool story though, bro.
|
On September 03 2010 03:22 theSAiNT wrote:Show nested quote +On September 03 2010 03:07 Sleight wrote:On September 03 2010 00:18 Sentient wrote:On September 03 2010 00:05 Sleight wrote:
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced. The trend holds over 160 players (the top 3 groups, more if you count the decline in Protoss). 20 players is a decent sample size, contrary to the random assertions otherwise. I will see if I have time tonight to actually do the numbers. You can draw statistics from these, and my intuition says there is a fair chunk of significance to them.And this deserves reposting: TT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected. Here is an example of a poster with only the most basic understanding of Statistics. I will just show how this is both wrong and embarrassing. To this first bolded bit: Twenty people cannot be significant in a population of 28K. In fact, as Buddhist (i believe) pointed out, the top 3 tiers amount to 160 people, which still remains just over 1 Percent of the population. 1% sample size is not a useful sample size in this kind of sampling analysis unless it is a truly random assortment. The term we are referring to by "sample size" is the "power" of said sample. When you have a small sample, you have less power unless the different you are looking for is HUGE. Actually being involved in professional population anaylses as part of my medical degree, I can say that this sample size LACKS SUFFICIENT POWER to obtain any result given the overall population. Since you deal in 'professional population analyses' in medicine, just out of interest, how many drugs passed by the FSA or indeed any other international body have been trialed on more than 1% of the population? Your conclusion is also wrong. You need sample variances to determine power. Sample size is insufficient. Notwithstanding that, nobody is actually claiming that the 'top X' is representative of the whole population. All the claims refer to balance within the 'top X'. In which case, the sample size is the whole population. Still, to make any claims about 'power' you need sample variances.
Yes there are people making that claim.
Edit - see the post just below this one.
|
Terrans imbalance has been "proven" by other means, these statistics are not going to make them any more imbalanced or give more weight to an already case closed argument.
Now...
Does this chart correlate with what we already know?
Is the Terran imbalance, which is already confirmed by Blizzard, the reason we have so many Terran at the top?
What will be interesting is if post patch 1.1 these charts change and show a different result.
This is just partial data for an overall experiment.
The experiment is "Is the most powerful race also the most common race at 1500+"
So far it looks that way, but we will need to have a different race be confirmed as imbalanced and run this test again.
Once we run this test numerous times with all the different imbalanced races over the course of many patches we can confirm or deny that these charts mean anything.
Right now it is only partial data that appears to correlate with imbalance, but doesn't truly mean that it does.
|
On September 03 2010 03:22 theSAiNT wrote:Show nested quote +On September 03 2010 03:07 Sleight wrote:On September 03 2010 00:18 Sentient wrote:On September 03 2010 00:05 Sleight wrote:
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced. The trend holds over 160 players (the top 3 groups, more if you count the decline in Protoss). 20 players is a decent sample size, contrary to the random assertions otherwise. I will see if I have time tonight to actually do the numbers. You can draw statistics from these, and my intuition says there is a fair chunk of significance to them.And this deserves reposting: TT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected. Here is an example of a poster with only the most basic understanding of Statistics. I will just show how this is both wrong and embarrassing. To this first bolded bit: Twenty people cannot be significant in a population of 28K. In fact, as Buddhist (i believe) pointed out, the top 3 tiers amount to 160 people, which still remains just over 1 Percent of the population. 1% sample size is not a useful sample size in this kind of sampling analysis unless it is a truly random assortment. The term we are referring to by "sample size" is the "power" of said sample. When you have a small sample, you have less power unless the different you are looking for is HUGE. Actually being involved in professional population anaylses as part of my medical degree, I can say that this sample size LACKS SUFFICIENT POWER to obtain any result given the overall population. Since you deal in 'professional population analyses' in medicine, just out of interest, how many drugs passed by the FSA or indeed any other international body have been trialed on more than 1% of the population? Your conclusion is also wrong. You need sample variances to determine power. Sample size is insufficient. Notwithstanding that, nobody is actually claiming that the 'top X' is representative of the whole population. All the claims refer to balance within the 'top X'. In which case, the sample size is the whole population. Still, to make any claims about 'power' you need sample variances.
You are an idiot. The reason drug trials, only phase 3 are truly random, are acceptable is a) MEDICAL NECESSSITY and b) TRULY RANDOM populations.
They succeed because of TRUE RANDOMNESS. They don't pick the tallest population, they don't pick white people, they just take PEOPLE. The population at the TOP OF A GAME cannot be random BY ITS NATURE.
EDIT;
If you separate a population BASED ON SKILL and then test those groups to determine SKILL, you will find out that, WOW, the better groups are better. If you separate groups on SKILL and try to determine something that SKILL is directly dependent on, it introduces BIA into a study. Biased data = biased conclusion.
|
On September 03 2010 03:14 Tray wrote:Show nested quote +On September 03 2010 02:56 Karkadinn wrote:On September 03 2010 02:51 Tray wrote:On September 03 2010 01:53 Karkadinn wrote:On September 03 2010 01:44 Tray wrote:On September 03 2010 01:40 Karkadinn wrote:On September 03 2010 01:06 Tray wrote:
Some of the first posters are correct. The data of the top-top diamonds is statistically insignificant because of the very small number of players at this level.
Basically the graph becomes less representative of reality the higher you go up the point score. It 'could' be accurate, but this data does not prove that, so the people QQing about Terran so OP should probably learn a little statistics. Doesn't matter that it's the "entire population." If that's your reasoning you have no clue what you're talking about here. There will always be a very small number of players at high levels. You're as good as saying that it's impossible to use statistics to measure anything significant, since the match-maker will by definition even out populations at more casual levels. At the top level, with small playerbase, yes it is impossible. That's exactly the point. You need to have a decent number of players before you're getting a clear view of the actual situation. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. You're taking the absolute extreme to support your point. Even if you just stop at 1200, the differences are sufficiently obvious. Trends among the top players are also important because one faction mirror matches do not make particularly good media entertainment for tournaments. Even in the 1200+ the stats are still probably not statistically relevant. As someone posted, that's less than 1.5% of all diamond players. I'm not gonna do the math, just trying to enlighten some people on stats here because there's a lot of sheep jumping to horribly retarded conclusions. Good luck convincing all the 'sheep' of your pov when your methodology consists of declaring all information beyond an arbitrary self-determined point completely meaningless. Who do I have to convince? The burden of proof is on you to prove to me that Terran is imbalanced because of this. Statistically, that has not been shown. 1200+ is just as arbitrary as any other cutoff. You really think diamonds 800-1200 should be ignored? I think Idra falls into that category... Seriously you people are too dumb to even discuss this topic with at an adult level. I'm guessing none of you have even taken a stats class, let alone graduated high school.
Have fun maintaining your facade of maturity while simultaneously insulting everyone who disagrees with you. If this is the level of civility I can expect when debating with you, I'm not interested in further interaction.
|
On September 03 2010 03:24 Tray wrote:Show nested quote +On September 03 2010 03:20 Bibdy wrote:
I imagine when a fanatical cult sect comes charging into his hometown in order to lynch all of the non-whites in the place, Tray will be standing there, arms crossed, telling them the burden of proof is on them to show that all the non-whites need to be lynched.
They're going to believe what they want to believe until you give them a good reason to believe otherwise. If you genuinely don't care what they believe, then why are you just shouting and screaming at people that they're wrong? If you do care, why are you too lazy to do the math?
That's just textbook trolling right there. No what I'm doing is telling people why they're wrong. What YOU'RE doing is trolling. You're great at it. You've been doing it since the beta in the bnet forums. I have a job, so I don't have all day to crunch numbers in excel to prove to you that your sample is statistically irrelevant. Go do a little google search on stats, confidence intervals, and the like. You might learn something. Cool story though, bro.
No, you're telling people they're wrong, retarded, stupid, haven't got a high school education and whatever other insults you thrown out prior to just the last page alone, without actually saying why.
Just a lot of "Wow, how could you POSSIBLY be this stupid? Me and my high horse are going to keep hanging out with you idiots, though".
Oh and you're the same Tray? Someone still needs to take a chill-pill I see. Remember how 'wrong' I was about Marauders?
|
I can't understand why people continue to make these efforts to show that terran is op or that terran is perfectly fine. You are not going to convince anyone who started out with a different opinion than you, and since you aren't trying to convince blizzard, I can't figure out what else you are trying to do. Every single "statistical" analysis gets called out for incorrect methods, incorrect assumptions, or some other failure.
I play zerg at the diamond level, and I feel that terran has too many options and zerg has too few options early on, but I can see myself convincing a single person who doesn't already think that way.
Thats my 15.5 cents.
|
Think of this graph as merely a peice of evidence. No one bit of evidence should be used to declare something like "imbalance" but taken in addition to other evidence people can start to make conclusions. For example, I have a friend who plays Terran, is missing some very basic understanding of the game, and has 1 build order no matter what the circumstance; 3 rax mm. I've tried teaching him 1-1-1 with very mixed results. He has ridden terran and mmm all the way to 500+ diamond with minimal effort. Combining that with the graph, my own experiences at ~900 diamond, and a basic understanding of the game, I can come to the conclusion that Terran currently is either easier to play well or allows you to play above your skill level, backed up by the trending scene in the data.
Now, I personally don't think this is a huge deal. I can play against Terran well, and I like the idea of small nerfs for specific things over time. But this is one piece of a volume of evidence, nothing more, nothing less.
|
On September 03 2010 03:07 Sleight wrote:Show nested quote +On September 03 2010 00:18 Sentient wrote:On September 03 2010 00:05 Sleight wrote:
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced. The trend holds over 160 players (the top 3 groups, more if you count the decline in Protoss). 20 players is a decent sample size, contrary to the random assertions otherwise. I will see if I have time tonight to actually do the numbers. You can draw statistics from these, and my intuition says there is a fair chunk of significance to them.And this deserves reposting: TT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected. Here is an example of a poster with only the most basic understanding of Statistics. I will just show how this is both wrong and embarrassing. To this first bolded bit: Twenty people cannot be significant in a population of 28K. In fact, as Buddhist (i believe) pointed out, the top 3 tiers amount to 160 people, which still remains just over 1 Percent of the population. 1% sample size is not a useful sample size in this kind of sampling analysis unless it is a truly random assortment. The term we are referring to by "sample size" is the "power" of said sample. When you have a small sample, you have less power unless the different you are looking for is HUGE. Actually being involved in professional population anaylses as part of my medical degree, I can say that this sample size LACKS SUFFICIENT POWER to obtain any result given the overall population. Furthermore, by simply picking the TOP you are eliminating a random element. This is stratified sampling and is inclined to a number of kinds of bias and is not viable unless you can demonstrate that the method is necessary to create a balanced population, This is not the case and is therefore not an acceptable sampling method. To the second boldded bit: Based on the graphs presented, of the population of diamond players > 600, Protoss is actually the most prevalent. In fact, a MAJORITY do NOT play Terran. A PLURALITY may play Terran, though not in this sampling, meaning the largest group less than 50%, but over 50% of players, which is the definition of majority, do not play Terran. Not in this graph, not anywhere. Just count them. So to say the number of players >1300 play Terran means nothing unless this is a sampling with adequate power and sample size to demonstrate statistical significance. I have already explained to you that it does not. Furthermore, you cannot assume TRENDS from a CROSS-SECTIONAL analysis. This type of study only can look at "prevalence" or the actual state of people at this moment. It cannot tell if people are moving up, down, or dying, for all it matters. You can only say "In the 20 person sample size at the highest tier, there are 12 Terran." Or, "in the top 3 tiers of 120 people, Terran comprises a larger than expected statistically significant portion by chi-squared analysis, WITHIN THIS POPULATION." If we try to compare the results to larger populations, we find that they lack statistical significance. Conclusion: Sorry, Sentient, one course in community college does not qualify you to be a statistical analyst. And you are wrong. A double whammy. Stop being thick and study harder.
sigh...
Imagine you want to examine the color of the ultra rare mexican swamp tree frog (mtf). Now because this frog is soooooo ultra rare you can look at the whole population of MSTFs. 60% are yellow, 30% are blue and 10% are red. You got 360 of these frogs and do a test to see the propability of the colors being randomly distributed. The tests says: No, with a propability of x% these numbers are not random. (see previous post). Congratulations, you found out something!
Now imagine people are coming and telling you that the MSTF only makes up much less than 1% of all the frogs and start talking about random samples blablabla. But you dont care. You said that your results are valid for this one type of frog (or only valid for all people >1200). And your results are pretty clear and significant. And it absolutely does not matter even the slightest how many other frogs there are. NOT EVEN A TINY BIT.
To go back to SC2: even if 200 million billion people were playing zerg in bronze would not mean that the data for the top players (>1200) is wrong. Fact is: there is a disproportional amount of terrans at the top.
Now if you think what would you expect if there was an imbalance: You would expect that the players for race x are shifted by y amount of points and this would be most apparent at the absolut top and the bottom. Hintedi hint hint.
But feel free to do another nice ad hominem and tell me that my university classes sucked and that I also have to study harder. Alternatively you could read about cognitive biases and discover, that people do not modify their views of the world to fit new information but rather fit the information into their views and beliefs. With hillarious results.
|
I want to see this graph not in % but in ACTUAL NUMBERS, the % can really screw with perception of what people see in a graph, actually all graphs can be easily tweaked just to show what people want to see. Ignoring the "terran is op omg" debate, I think its important we see the actual figures behind these graphs.
|
On September 03 2010 03:28 Bibdy wrote:Show nested quote +On September 03 2010 03:24 Tray wrote:On September 03 2010 03:20 Bibdy wrote:
I imagine when a fanatical cult sect comes charging into his hometown in order to lynch all of the non-whites in the place, Tray will be standing there, arms crossed, telling them the burden of proof is on them to show that all the non-whites need to be lynched.
They're going to believe what they want to believe until you give them a good reason to believe otherwise. If you genuinely don't care what they believe, then why are you just shouting and screaming at people that they're wrong? If you do care, why are you too lazy to do the math?
That's just textbook trolling right there. No what I'm doing is telling people why they're wrong. What YOU'RE doing is trolling. You're great at it. You've been doing it since the beta in the bnet forums. I have a job, so I don't have all day to crunch numbers in excel to prove to you that your sample is statistically irrelevant. Go do a little google search on stats, confidence intervals, and the like. You might learn something. Cool story though, bro. No, you're telling people they're wrong, retarded, stupid, haven't got a high school education and whatever other insults you thrown out prior to just the last page alone, without actually saying why. Just a lot of "Wow, how could you POSSIBLY be this stupid? Me and my high horse are going to keep hanging out with you idiots, though". Oh and you're the same Tray? Someone still needs to take a chill-pill I see. Remember how 'wrong' I was about Marauders?
"Anti-intellectualism has been a constant thread winding its way through our political and cultural life, nurtured by the false notion that democracy means that 'my ignorance is just as good as your knowledge.'" - Isaac Asimov"
I back up my calling you stupid because you are and don't understand stats, and I explain to you why your analysis is incorrect. You counter with hyperbole and consider that a logically sound argument. Only one of us is a troll, guess which one?
Also you were wrong about marauders. Notice how they haven't been nerfed? I mean, are you serious? You're proving my points for me.
|
On September 03 2010 03:26 Sleight wrote:Show nested quote +On September 03 2010 03:22 theSAiNT wrote:On September 03 2010 03:07 Sleight wrote:On September 03 2010 00:18 Sentient wrote:On September 03 2010 00:05 Sleight wrote:
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced. The trend holds over 160 players (the top 3 groups, more if you count the decline in Protoss). 20 players is a decent sample size, contrary to the random assertions otherwise. I will see if I have time tonight to actually do the numbers. You can draw statistics from these, and my intuition says there is a fair chunk of significance to them.And this deserves reposting: TT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected. Here is an example of a poster with only the most basic understanding of Statistics. I will just show how this is both wrong and embarrassing. To this first bolded bit: Twenty people cannot be significant in a population of 28K. In fact, as Buddhist (i believe) pointed out, the top 3 tiers amount to 160 people, which still remains just over 1 Percent of the population. 1% sample size is not a useful sample size in this kind of sampling analysis unless it is a truly random assortment. The term we are referring to by "sample size" is the "power" of said sample. When you have a small sample, you have less power unless the different you are looking for is HUGE. Actually being involved in professional population anaylses as part of my medical degree, I can say that this sample size LACKS SUFFICIENT POWER to obtain any result given the overall population. Since you deal in 'professional population analyses' in medicine, just out of interest, how many drugs passed by the FSA or indeed any other international body have been trialed on more than 1% of the population? Your conclusion is also wrong. You need sample variances to determine power. Sample size is insufficient. Notwithstanding that, nobody is actually claiming that the 'top X' is representative of the whole population. All the claims refer to balance within the 'top X'. In which case, the sample size is the whole population. Still, to make any claims about 'power' you need sample variances. You are an idiot. The reason drug trials, only phase 3 are truly random, are acceptable is a) MEDICAL NECESSSITY and b) TRULY RANDOM populations. They succeed because of TRUE RANDOMNESS. They don't pick the tallest population, they don't pick white people, they just take PEOPLE. The population at the TOP OF A GAME cannot be random BY ITS NATURE.
Your initial emphasis was on the small size of the sample.
Nobody is actually saying that the top percentage are randomly selected. Nobody is saying they are representative of the whole population. The claim is that 'at the top level of play, there is evidence of imbalance'.
And to answer my question, no. Drug trials are often passed with sample sizes less than 50. Not anywhere even close to 1% of the population. 1% of a population is a lot.
|
On September 03 2010 03:28 icedragon wrote:
I can't understand why people continue to make these efforts to show that terran is op or that terran is perfectly fine. You are not going to convince anyone who started out with a different opinion than you, and since you aren't trying to convince blizzard, I can't figure out what else you are trying to do. Every single "statistical" analysis gets called out for incorrect methods, incorrect assumptions, or some other failure.
I play zerg at the diamond level, and I feel that terran has too many options and zerg has too few options early on, but I can see myself convincing a single person who doesn't already think that way.
Thats my 15.5 cents.
In theory, people are supposed to start out from a point of being persuadable and are then persuaded by the presented arguments on either side. That's, depressingly, just theory. In fact, most people will cement their loyalty to an ideology or faction and will refuse to be persuaded from it regardless of evidence. So what complaining comes down to online is really just plain venting and attempting to get emotional validation for what you feel are your grievances. In a perfect world, maybe someone could actually convince someone else of something with a logical argument, but in this one... meh... it's mostly people yelling at each other. And we keep doing it anyway, because we need to vent, and we need validation. This is readily observable in gaming communities, but you can see it in anything, especially politics and theology. I'm not sure if there's any real way to 'fix' this. If you ban people for being close-minded, you'd end up banning everyone eventually.
|
On September 03 2010 03:33 theSAiNT wrote:Show nested quote +On September 03 2010 03:26 Sleight wrote:On September 03 2010 03:22 theSAiNT wrote:On September 03 2010 03:07 Sleight wrote:On September 03 2010 00:18 Sentient wrote:On September 03 2010 00:05 Sleight wrote:
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced. The trend holds over 160 players (the top 3 groups, more if you count the decline in Protoss). 20 players is a decent sample size, contrary to the random assertions otherwise. I will see if I have time tonight to actually do the numbers. You can draw statistics from these, and my intuition says there is a fair chunk of significance to them.And this deserves reposting: TT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected. Here is an example of a poster with only the most basic understanding of Statistics. I will just show how this is both wrong and embarrassing. To this first bolded bit: Twenty people cannot be significant in a population of 28K. In fact, as Buddhist (i believe) pointed out, the top 3 tiers amount to 160 people, which still remains just over 1 Percent of the population. 1% sample size is not a useful sample size in this kind of sampling analysis unless it is a truly random assortment. The term we are referring to by "sample size" is the "power" of said sample. When you have a small sample, you have less power unless the different you are looking for is HUGE. Actually being involved in professional population anaylses as part of my medical degree, I can say that this sample size LACKS SUFFICIENT POWER to obtain any result given the overall population. Since you deal in 'professional population analyses' in medicine, just out of interest, how many drugs passed by the FSA or indeed any other international body have been trialed on more than 1% of the population? Your conclusion is also wrong. You need sample variances to determine power. Sample size is insufficient. Notwithstanding that, nobody is actually claiming that the 'top X' is representative of the whole population. All the claims refer to balance within the 'top X'. In which case, the sample size is the whole population. Still, to make any claims about 'power' you need sample variances. You are an idiot. The reason drug trials, only phase 3 are truly random, are acceptable is a) MEDICAL NECESSSITY and b) TRULY RANDOM populations. They succeed because of TRUE RANDOMNESS. They don't pick the tallest population, they don't pick white people, they just take PEOPLE. The population at the TOP OF A GAME cannot be random BY ITS NATURE. Your initial emphasis was on the small size of the sample. Nobody is actually saying that the top percentage are randomly selected. Nobody is saying they are representative of the whole population. The claim is that 'at the top level of play, there is evidence of imbalance'. And to answer my question, no. Drug trials are often passed with sample sizes less than 50. Not anywhere even close to 1% of the population. 1% of a population is a lot.
Yes they are. On this page of the thread in fact. Read it.
And no the claim that at the top levels of play that there is some imbalance is exactly what we're proving is not true. This is not evidence. That's the entire point. Try to keep up.
|
On September 03 2010 03:32 Tray wrote:Show nested quote +On September 03 2010 03:28 Bibdy wrote:On September 03 2010 03:24 Tray wrote:On September 03 2010 03:20 Bibdy wrote:
I imagine when a fanatical cult sect comes charging into his hometown in order to lynch all of the non-whites in the place, Tray will be standing there, arms crossed, telling them the burden of proof is on them to show that all the non-whites need to be lynched.
They're going to believe what they want to believe until you give them a good reason to believe otherwise. If you genuinely don't care what they believe, then why are you just shouting and screaming at people that they're wrong? If you do care, why are you too lazy to do the math?
That's just textbook trolling right there. No what I'm doing is telling people why they're wrong. What YOU'RE doing is trolling. You're great at it. You've been doing it since the beta in the bnet forums. I have a job, so I don't have all day to crunch numbers in excel to prove to you that your sample is statistically irrelevant. Go do a little google search on stats, confidence intervals, and the like. You might learn something. Cool story though, bro. No, you're telling people they're wrong, retarded, stupid, haven't got a high school education and whatever other insults you thrown out prior to just the last page alone, without actually saying why. Just a lot of "Wow, how could you POSSIBLY be this stupid? Me and my high horse are going to keep hanging out with you idiots, though". Oh and you're the same Tray? Someone still needs to take a chill-pill I see. Remember how 'wrong' I was about Marauders? "Anti-intellectualism has been a constant thread winding its way through our political and cultural life, nurtured by the false notion that democracy means that 'my ignorance is just as good as your knowledge.'" - Isaac Asimov" I back up my calling you stupid because you are and don't understand stats, and I explain to you why your analysis is incorrect. You counter with hyperbole and consider that a logically sound argument. Only one of us is a troll, guess which one? Also you were wrong about marauders. Notice how they haven't been nerfed? I mean, are you serious? You're proving my points for me.
“When an argument flares up, the wise man quenches it with silence” - Anonymous
See, I can google a quote and sound smart, too.
Oh and 'cept for that whole Conc Shell upgrade thing, yeah. No nerfs at all.
|
On September 03 2010 03:32 fert wrote:
I want to see this graph not in % but in ACTUAL NUMBERS, the % can really screw with perception of what people see in a graph, actually all graphs can be easily tweaked just to show what people want to see. Ignoring the "terran is op omg" debate, I think its important we see the actual figures behind these graphs.
I believe there is actually a link for what you're asking for in the original post.
|
On September 03 2010 03:12 Tray wrote:Show nested quote +On September 03 2010 03:08 Winter_mute wrote:On September 03 2010 02:53 Tray wrote:On September 03 2010 01:58 Fraud wrote:I enjoy the references by people of the "sample size being too small", when the sample size is equal to 100% of the population. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. Everyone was previously saying we should only be balancing at the top of the game, as if you're not near the top, you can advance by getting better. Looking at 1200+ Diamond, a clear pattern is emerging. Terran's are dominating the top 400 players. That being said, this graph proves what has already been said multiple times, Terrans have an advantage at the top and Zerg is weak. That's why Blizzard is releasing Patch 1.1 This is not true and the person who posted this is not very smart. Please don't post on statistics if you don't understand it. I wrote it before, but I guess I have to write it again: On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200. Since you obviously understand statistics, you can check for yourself and provide your numbers. And of course you are allowed to normalize the null hypothesis to account for all these players who just play terran for the looks or because they played them in the campaign. Just tell us the corrected null hypothesis numbers. You people are so retarded. You're not even accounting for the number of people that play terran relative to the other races. Do you really think your stats are even close to relevant? Wow. Go back to school.
You Sir are the retard, if you think you can call all the people disagreeing with you retarded. There are two questions to be asked. First: "Why are so many people favoring T over Z?". Second which partly answers the first: "Why are they performing better?"
|
On September 03 2010 03:35 Tray wrote:Show nested quote +On September 03 2010 03:33 theSAiNT wrote:On September 03 2010 03:26 Sleight wrote:On September 03 2010 03:22 theSAiNT wrote:On September 03 2010 03:07 Sleight wrote:On September 03 2010 00:18 Sentient wrote:On September 03 2010 00:05 Sleight wrote:
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced. The trend holds over 160 players (the top 3 groups, more if you count the decline in Protoss). 20 players is a decent sample size, contrary to the random assertions otherwise. I will see if I have time tonight to actually do the numbers. You can draw statistics from these, and my intuition says there is a fair chunk of significance to them.And this deserves reposting: TT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected. Here is an example of a poster with only the most basic understanding of Statistics. I will just show how this is both wrong and embarrassing. To this first bolded bit: Twenty people cannot be significant in a population of 28K. In fact, as Buddhist (i believe) pointed out, the top 3 tiers amount to 160 people, which still remains just over 1 Percent of the population. 1% sample size is not a useful sample size in this kind of sampling analysis unless it is a truly random assortment. The term we are referring to by "sample size" is the "power" of said sample. When you have a small sample, you have less power unless the different you are looking for is HUGE. Actually being involved in professional population anaylses as part of my medical degree, I can say that this sample size LACKS SUFFICIENT POWER to obtain any result given the overall population. Since you deal in 'professional population analyses' in medicine, just out of interest, how many drugs passed by the FSA or indeed any other international body have been trialed on more than 1% of the population? Your conclusion is also wrong. You need sample variances to determine power. Sample size is insufficient. Notwithstanding that, nobody is actually claiming that the 'top X' is representative of the whole population. All the claims refer to balance within the 'top X'. In which case, the sample size is the whole population. Still, to make any claims about 'power' you need sample variances. You are an idiot. The reason drug trials, only phase 3 are truly random, are acceptable is a) MEDICAL NECESSSITY and b) TRULY RANDOM populations. They succeed because of TRUE RANDOMNESS. They don't pick the tallest population, they don't pick white people, they just take PEOPLE. The population at the TOP OF A GAME cannot be random BY ITS NATURE. Your initial emphasis was on the small size of the sample. Nobody is actually saying that the top percentage are randomly selected. Nobody is saying they are representative of the whole population. The claim is that 'at the top level of play, there is evidence of imbalance'. And to answer my question, no. Drug trials are often passed with sample sizes less than 50. Not anywhere even close to 1% of the population. 1% of a population is a lot. Yes they are. On this page of the thread in fact. Read it. And no the claim that at the top levels of play that there is some imbalance is exactly what we're proving is not true. This is not evidence. That's the entire point. Try to keep up.
You can put away the patronizing tone. Even if they are, they're not reaching that claim with a 'statistical' statement that the top X% is a randomly selected, representative sample of the whole population which is the basis of your refutation.
For example:
On September 03 2010 03:32 Winter_mute wrote:
Now if you think what would you expect if there was an imbalance: You would expect that the players for race x are shifted by y amount of points and this would be most apparent at the absolut top and the bottom. Hintedi hint hint.
|
Was TL this way the whole time? Or is it just the newfags trolling, hating and spamming these forums about imbalance when they know shit about the game? yo
|
On September 03 2010 03:39 Perkins1752 wrote:Show nested quote +On September 03 2010 03:12 Tray wrote:On September 03 2010 03:08 Winter_mute wrote:On September 03 2010 02:53 Tray wrote:On September 03 2010 01:58 Fraud wrote:I enjoy the references by people of the "sample size being too small", when the sample size is equal to 100% of the population. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. Everyone was previously saying we should only be balancing at the top of the game, as if you're not near the top, you can advance by getting better. Looking at 1200+ Diamond, a clear pattern is emerging. Terran's are dominating the top 400 players. That being said, this graph proves what has already been said multiple times, Terrans have an advantage at the top and Zerg is weak. That's why Blizzard is releasing Patch 1.1 This is not true and the person who posted this is not very smart. Please don't post on statistics if you don't understand it. I wrote it before, but I guess I have to write it again: On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200. Since you obviously understand statistics, you can check for yourself and provide your numbers. And of course you are allowed to normalize the null hypothesis to account for all these players who just play terran for the looks or because they played them in the campaign. Just tell us the corrected null hypothesis numbers. You people are so retarded. You're not even accounting for the number of people that play terran relative to the other races. Do you really think your stats are even close to relevant? Wow. Go back to school. You Sir are the retard, if you think you can call all the people disagreeing with you retarded. There are two questions to be asked. First: "Why are so many people favoring T over Z?". Second which partly answers the first: "Why are they performing better?"
I am? Look in the mirror. People probably play Terran because the campaign is terran. A hugely disproportionate number of NA players play Terran. To find a true statistical inaccuracy you would have to compare the % of players who play terran to the % of Terran players in the top X bracket. In this case you don't have enough of a sample to prove any disparity.
To answer "Why are they performing better?" They aren't. Pretty simple.
|
Doesn't this about match the overall spread of the races until about 1200? Can't find the numbers, atm.
|
On September 03 2010 03:32 Winter_mute wrote:Show nested quote +On September 03 2010 03:07 Sleight wrote:On September 03 2010 00:18 Sentient wrote:On September 03 2010 00:05 Sleight wrote:
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced. The trend holds over 160 players (the top 3 groups, more if you count the decline in Protoss). 20 players is a decent sample size, contrary to the random assertions otherwise. I will see if I have time tonight to actually do the numbers. You can draw statistics from these, and my intuition says there is a fair chunk of significance to them.And this deserves reposting: TT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected. Here is an example of a poster with only the most basic understanding of Statistics. I will just show how this is both wrong and embarrassing. To this first bolded bit: Twenty people cannot be significant in a population of 28K. In fact, as Buddhist (i believe) pointed out, the top 3 tiers amount to 160 people, which still remains just over 1 Percent of the population. 1% sample size is not a useful sample size in this kind of sampling analysis unless it is a truly random assortment. The term we are referring to by "sample size" is the "power" of said sample. When you have a small sample, you have less power unless the different you are looking for is HUGE. Actually being involved in professional population anaylses as part of my medical degree, I can say that this sample size LACKS SUFFICIENT POWER to obtain any result given the overall population. Furthermore, by simply picking the TOP you are eliminating a random element. This is stratified sampling and is inclined to a number of kinds of bias and is not viable unless you can demonstrate that the method is necessary to create a balanced population, This is not the case and is therefore not an acceptable sampling method. To the second boldded bit: Based on the graphs presented, of the population of diamond players > 600, Protoss is actually the most prevalent. In fact, a MAJORITY do NOT play Terran. A PLURALITY may play Terran, though not in this sampling, meaning the largest group less than 50%, but over 50% of players, which is the definition of majority, do not play Terran. Not in this graph, not anywhere. Just count them. So to say the number of players >1300 play Terran means nothing unless this is a sampling with adequate power and sample size to demonstrate statistical significance. I have already explained to you that it does not. Furthermore, you cannot assume TRENDS from a CROSS-SECTIONAL analysis. This type of study only can look at "prevalence" or the actual state of people at this moment. It cannot tell if people are moving up, down, or dying, for all it matters. You can only say "In the 20 person sample size at the highest tier, there are 12 Terran." Or, "in the top 3 tiers of 120 people, Terran comprises a larger than expected statistically significant portion by chi-squared analysis, WITHIN THIS POPULATION." If we try to compare the results to larger populations, we find that they lack statistical significance. Conclusion: Sorry, Sentient, one course in community college does not qualify you to be a statistical analyst. And you are wrong. A double whammy. Stop being thick and study harder. sigh... Imagine you want to examine the color of the ultra rare mexican swamp tree frog (mtf). Now because this frog is soooooo ultra rare you can look at the whole population of MSTFs. 60% are yellow, 30% are blue and 10% are red. You got 360 of these frogs and do a test to see the propability of the colors being randomly distributed. The tests says: No, with a propability of x% these numbers are not random. (see previous post). Congratulations, you found out something! Now imagine people are coming and telling you that the MSTF only makes up much less than 1% of all the frogs and start talking about random samples blablabla. But you dont care. You said that your results are valid for this one type of frog (or only valid for all people >1200). And your results are pretty clear and significant. And it absolutely does not matter even the slightest how many other frogs there are. NOT EVEN A TINY BIT. To go back to SC2: even if 200 million billion people were playing zerg in bronze would not mean that the data for the top players (>1200) is wrong. Fact is: there is a disproportional amount of terrans at the top. Now if you think what would you expect if there was an imbalance: You would expect that the players for race x are shifted by y amount of points and this would be most apparent at the absolut top and the bottom. Hintedi hint hint. But feel free to do another nice ad hominem and tell me that my university classes sucked and that I also have to study harder. Alternatively you could read about cognitive biases and discover, that people do not modify their views of the world to fit new information but rather fit the information into their views and beliefs. With hillarious results.
Agreed with this guy.
Tray you are embarrassing yourself.
|
On September 03 2010 03:43 Tray wrote:Show nested quote +On September 03 2010 03:39 Perkins1752 wrote:On September 03 2010 03:12 Tray wrote:On September 03 2010 03:08 Winter_mute wrote:On September 03 2010 02:53 Tray wrote:On September 03 2010 01:58 Fraud wrote:I enjoy the references by people of the "sample size being too small", when the sample size is equal to 100% of the population. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. Everyone was previously saying we should only be balancing at the top of the game, as if you're not near the top, you can advance by getting better. Looking at 1200+ Diamond, a clear pattern is emerging. Terran's are dominating the top 400 players. That being said, this graph proves what has already been said multiple times, Terrans have an advantage at the top and Zerg is weak. That's why Blizzard is releasing Patch 1.1 This is not true and the person who posted this is not very smart. Please don't post on statistics if you don't understand it. I wrote it before, but I guess I have to write it again: On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200. Since you obviously understand statistics, you can check for yourself and provide your numbers. And of course you are allowed to normalize the null hypothesis to account for all these players who just play terran for the looks or because they played them in the campaign. Just tell us the corrected null hypothesis numbers. You people are so retarded. You're not even accounting for the number of people that play terran relative to the other races. Do you really think your stats are even close to relevant? Wow. Go back to school. You Sir are the retard, if you think you can call all the people disagreeing with you retarded. There are two questions to be asked. First: "Why are so many people favoring T over Z?". Second which partly answers the first: "Why are they performing better?" I am? Look in the mirror. People probably play Terran because the campaign is terran. A hugely disproportionate number of NA players play Terran. To find a true statistical inaccuracy you would have to compare the % of players who play terran to the % of Terran players in the top X bracket. In this case you don't have enough of a sample to prove any disparity. To answer "Why are they performing better?" They aren't. Pretty simple.
I don't particularly care, but I'm pretty sure that more people play protoss than terran. You should probably check that out before arguing about it (even with the word PROBABLY in front of the argument)
|
On September 03 2010 03:55 ploy wrote:Show nested quote +On September 03 2010 03:32 Winter_mute wrote:On September 03 2010 03:07 Sleight wrote:On September 03 2010 00:18 Sentient wrote:On September 03 2010 00:05 Sleight wrote:
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced. The trend holds over 160 players (the top 3 groups, more if you count the decline in Protoss). 20 players is a decent sample size, contrary to the random assertions otherwise. I will see if I have time tonight to actually do the numbers. You can draw statistics from these, and my intuition says there is a fair chunk of significance to them.And this deserves reposting: TT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected. Here is an example of a poster with only the most basic understanding of Statistics. I will just show how this is both wrong and embarrassing. To this first bolded bit: Twenty people cannot be significant in a population of 28K. In fact, as Buddhist (i believe) pointed out, the top 3 tiers amount to 160 people, which still remains just over 1 Percent of the population. 1% sample size is not a useful sample size in this kind of sampling analysis unless it is a truly random assortment. The term we are referring to by "sample size" is the "power" of said sample. When you have a small sample, you have less power unless the different you are looking for is HUGE. Actually being involved in professional population anaylses as part of my medical degree, I can say that this sample size LACKS SUFFICIENT POWER to obtain any result given the overall population. Furthermore, by simply picking the TOP you are eliminating a random element. This is stratified sampling and is inclined to a number of kinds of bias and is not viable unless you can demonstrate that the method is necessary to create a balanced population, This is not the case and is therefore not an acceptable sampling method. To the second boldded bit: Based on the graphs presented, of the population of diamond players > 600, Protoss is actually the most prevalent. In fact, a MAJORITY do NOT play Terran. A PLURALITY may play Terran, though not in this sampling, meaning the largest group less than 50%, but over 50% of players, which is the definition of majority, do not play Terran. Not in this graph, not anywhere. Just count them. So to say the number of players >1300 play Terran means nothing unless this is a sampling with adequate power and sample size to demonstrate statistical significance. I have already explained to you that it does not. Furthermore, you cannot assume TRENDS from a CROSS-SECTIONAL analysis. This type of study only can look at "prevalence" or the actual state of people at this moment. It cannot tell if people are moving up, down, or dying, for all it matters. You can only say "In the 20 person sample size at the highest tier, there are 12 Terran." Or, "in the top 3 tiers of 120 people, Terran comprises a larger than expected statistically significant portion by chi-squared analysis, WITHIN THIS POPULATION." If we try to compare the results to larger populations, we find that they lack statistical significance. Conclusion: Sorry, Sentient, one course in community college does not qualify you to be a statistical analyst. And you are wrong. A double whammy. Stop being thick and study harder. sigh... Imagine you want to examine the color of the ultra rare mexican swamp tree frog (mtf). Now because this frog is soooooo ultra rare you can look at the whole population of MSTFs. 60% are yellow, 30% are blue and 10% are red. You got 360 of these frogs and do a test to see the propability of the colors being randomly distributed. The tests says: No, with a propability of x% these numbers are not random. (see previous post). Congratulations, you found out something! Now imagine people are coming and telling you that the MSTF only makes up much less than 1% of all the frogs and start talking about random samples blablabla. But you dont care. You said that your results are valid for this one type of frog (or only valid for all people >1200). And your results are pretty clear and significant. And it absolutely does not matter even the slightest how many other frogs there are. NOT EVEN A TINY BIT. To go back to SC2: even if 200 million billion people were playing zerg in bronze would not mean that the data for the top players (>1200) is wrong. Fact is: there is a disproportional amount of terrans at the top. Now if you think what would you expect if there was an imbalance: You would expect that the players for race x are shifted by y amount of points and this would be most apparent at the absolut top and the bottom. Hintedi hint hint. But feel free to do another nice ad hominem and tell me that my university classes sucked and that I also have to study harder. Alternatively you could read about cognitive biases and discover, that people do not modify their views of the world to fit new information but rather fit the information into their views and beliefs. With hillarious results. Agreed with this guy. Tray you are embarrassing yourself.
Haha no I'm not. You are. So is the guy you quoted. Not the same situation at all.
|
Ouch, 60% at high end? What. That's pretty dumb.
|
On September 03 2010 03:59 Tray wrote:Show nested quote +On September 03 2010 03:55 ploy wrote:On September 03 2010 03:32 Winter_mute wrote:On September 03 2010 03:07 Sleight wrote:On September 03 2010 00:18 Sentient wrote:On September 03 2010 00:05 Sleight wrote:
Hey y'all...
Look at the graph again and the n values at the bottom of each graph. The rightmost bar as TWENTY people in it. This will barely be statistically significant when referencing a bigger population that is known not to be perfectly balanced. The trend holds over 160 players (the top 3 groups, more if you count the decline in Protoss). 20 players is a decent sample size, contrary to the random assertions otherwise. I will see if I have time tonight to actually do the numbers. You can draw statistics from these, and my intuition says there is a fair chunk of significance to them.And this deserves reposting: TT: People who don't know statistics throwing around jargon like 'sample size'.
"Terran players make up the majority of 1300+ Diamond ladder players" - FACT. This is a population census. It is fully comprehensive in what it measures. There is no confidence to consider. These are the exact numbers for the moment in time when they were collected. Here is an example of a poster with only the most basic understanding of Statistics. I will just show how this is both wrong and embarrassing. To this first bolded bit: Twenty people cannot be significant in a population of 28K. In fact, as Buddhist (i believe) pointed out, the top 3 tiers amount to 160 people, which still remains just over 1 Percent of the population. 1% sample size is not a useful sample size in this kind of sampling analysis unless it is a truly random assortment. The term we are referring to by "sample size" is the "power" of said sample. When you have a small sample, you have less power unless the different you are looking for is HUGE. Actually being involved in professional population anaylses as part of my medical degree, I can say that this sample size LACKS SUFFICIENT POWER to obtain any result given the overall population. Furthermore, by simply picking the TOP you are eliminating a random element. This is stratified sampling and is inclined to a number of kinds of bias and is not viable unless you can demonstrate that the method is necessary to create a balanced population, This is not the case and is therefore not an acceptable sampling method. To the second boldded bit: Based on the graphs presented, of the population of diamond players > 600, Protoss is actually the most prevalent. In fact, a MAJORITY do NOT play Terran. A PLURALITY may play Terran, though not in this sampling, meaning the largest group less than 50%, but over 50% of players, which is the definition of majority, do not play Terran. Not in this graph, not anywhere. Just count them. So to say the number of players >1300 play Terran means nothing unless this is a sampling with adequate power and sample size to demonstrate statistical significance. I have already explained to you that it does not. Furthermore, you cannot assume TRENDS from a CROSS-SECTIONAL analysis. This type of study only can look at "prevalence" or the actual state of people at this moment. It cannot tell if people are moving up, down, or dying, for all it matters. You can only say "In the 20 person sample size at the highest tier, there are 12 Terran." Or, "in the top 3 tiers of 120 people, Terran comprises a larger than expected statistically significant portion by chi-squared analysis, WITHIN THIS POPULATION." If we try to compare the results to larger populations, we find that they lack statistical significance. Conclusion: Sorry, Sentient, one course in community college does not qualify you to be a statistical analyst. And you are wrong. A double whammy. Stop being thick and study harder. sigh... Imagine you want to examine the color of the ultra rare mexican swamp tree frog (mtf). Now because this frog is soooooo ultra rare you can look at the whole population of MSTFs. 60% are yellow, 30% are blue and 10% are red. You got 360 of these frogs and do a test to see the propability of the colors being randomly distributed. The tests says: No, with a propability of x% these numbers are not random. (see previous post). Congratulations, you found out something! Now imagine people are coming and telling you that the MSTF only makes up much less than 1% of all the frogs and start talking about random samples blablabla. But you dont care. You said that your results are valid for this one type of frog (or only valid for all people >1200). And your results are pretty clear and significant. And it absolutely does not matter even the slightest how many other frogs there are. NOT EVEN A TINY BIT. To go back to SC2: even if 200 million billion people were playing zerg in bronze would not mean that the data for the top players (>1200) is wrong. Fact is: there is a disproportional amount of terrans at the top. Now if you think what would you expect if there was an imbalance: You would expect that the players for race x are shifted by y amount of points and this would be most apparent at the absolut top and the bottom. Hintedi hint hint. But feel free to do another nice ad hominem and tell me that my university classes sucked and that I also have to study harder. Alternatively you could read about cognitive biases and discover, that people do not modify their views of the world to fit new information but rather fit the information into their views and beliefs. With hillarious results. Agreed with this guy. Tray you are embarrassing yourself. Haha no I'm not. You are. So is the guy you quoted. Not the same situation at all.
Embarassment is a social situation and/or a mental state based on a view of reality. Neither of you can decide that the other is embarrassing themselves.
|
On September 03 2010 03:43 Tray wrote:Show nested quote +On September 03 2010 03:39 Perkins1752 wrote:On September 03 2010 03:12 Tray wrote:On September 03 2010 03:08 Winter_mute wrote:On September 03 2010 02:53 Tray wrote:On September 03 2010 01:58 Fraud wrote:I enjoy the references by people of the "sample size being too small", when the sample size is equal to 100% of the population. For example there are only 25 players in the WORLD over 1500 points. If you think that's a sufficient number to base balance on because 11 of those players happen to play terran, then you have a very elementary understanding of statistical analysis. Everyone was previously saying we should only be balancing at the top of the game, as if you're not near the top, you can advance by getting better. Looking at 1200+ Diamond, a clear pattern is emerging. Terran's are dominating the top 400 players. That being said, this graph proves what has already been said multiple times, Terrans have an advantage at the top and Zerg is weak. That's why Blizzard is releasing Patch 1.1 This is not true and the person who posted this is not very smart. Please don't post on statistics if you don't understand it. I wrote it before, but I guess I have to write it again: On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200. Since you obviously understand statistics, you can check for yourself and provide your numbers. And of course you are allowed to normalize the null hypothesis to account for all these players who just play terran for the looks or because they played them in the campaign. Just tell us the corrected null hypothesis numbers. You people are so retarded. You're not even accounting for the number of people that play terran relative to the other races. Do you really think your stats are even close to relevant? Wow. Go back to school. You Sir are the retard, if you think you can call all the people disagreeing with you retarded. There are two questions to be asked. First: "Why are so many people favoring T over Z?". Second which partly answers the first: "Why are they performing better?" I am? Look in the mirror. People probably play Terran because the campaign is terran. A hugely disproportionate number of NA players play Terran. To find a true statistical inaccuracy you would have to compare the % of players who play terran to the % of Terran players in the top X bracket. In this case you don't have enough of a sample to prove any disparity. To answer "Why are they performing better?" They aren't. Pretty simple.
Dude everybody knows you are wrong. You don't give any evidence, yet you are insulting people. Weirdo.
|
The trend starting at 1200+ (180 pop) of increasing terran dominance flows very nicely right up to 1500+ (20 pop). This heavily supports the credibility of the 1500+ data, even though the population is lower.
Also, the growth of terran dominance from 1200 to 1500+ fits perfectly with how automatic match making gradually loses it's leveling effect when nearing the extremes.
Sorry if this was already mentioned.
|
On September 02 2010 07:46 cup of joe wrote:
it means absolutely nothing because the sample size is tiny at the right end of the graph
sample size is close to 100% lol...
|
yes indeed, cognitive bias in its finest, makes for very entertaining read
|
This actually makes me want to play Zerg. =P
|
The amount of people on here that don't understand basic statistics is astounding.
|
This is basically just 3 different 1 tail distributions that have an offset because of what I'll kindly call "racial differences in overall level of play".
|
is there some kind of international prize for the most hilarious statistics argument? This thread would win that without question.
|
And again, the numbers speak the truth, though the graph itself is terrible.
Unless you're playing at a very high level, blame your own fail rather than imbalance Z/P/T for your loss. If you are playing at a high level, it will be very exciting to see the effect of the next patch on balance there.
Regarding the graph... please draw it to scale. It is VERY misleading. I feel like the number of each bracket is in fine print... oh wait-.-
|
On September 03 2010 05:02 Emperor_Earth wrote:
And again, the numbers speak the truth, though the graph itself is terrible.
Unless you're playing at a very high level, blame your own fail rather than imbalance Z/P/T for your loss. If you are playing at a high level, it will be very exciting to see the effect of the next patch on balance there.
Regarding the graph... please draw it to scale. It is VERY misleading. I feel like the number of each bracket is in fine print... oh wait-.-
You want to see it to scale such that the bars on the left reach the top, and the bars on the right are about a pixel tall?
They're written by % with the axes clearly labeled. Just clicking on the image gives you a larger one to look at if you have to squint to read it.
|
On September 03 2010 03:32 Winter_mute wrote:
Imagine you want to examine the color of the ultra rare mexican swamp tree frog (mtf). Now because this frog is soooooo ultra rare you can look at the whole population of MSTFs. 60% are yellow, 30% are blue and 10% are red. You got 360 of these frogs and do a test to see the propability of the colors being randomly distributed. The tests says: No, with a propability of x% these numbers are not random. (see previous post). Congratulations, you found out something!
Now imagine people are coming and telling you that the MSTF only makes up much less than 1% of all the frogs and start talking about random samples blablabla. But you dont care. You said that your results are valid for this one type of frog (or only valid for all people >1200). And your results are pretty clear and significant. And it absolutely does not matter even the slightest how many other frogs there are. NOT EVEN A TINY BIT.
To go back to SC2: even if 200 million billion people were playing zerg in bronze would not mean that the data for the top players (>1200) is wrong. Fact is: there is a disproportional amount of terrans at the top.
Now if you think what would you expect if there was an imbalance: You would expect that the players for race x are shifted by y amount of points and this would be most apparent at the absolut top and the bottom. Hintedi hint hint.
But feel free to do another nice ad hominem and tell me that my university classes sucked and that I also have to study harder. Alternatively you could read about cognitive biases and discover, that people do not modify their views of the world to fit new information but rather fit the information into their views and beliefs. With hillarious results.
Great, this is an example of the misunderstanding.
Everything posted about the Tree Frog are dead on. Absolutely, if this was a random sampling of a rare population, this could not be more correct.
Here is what I said and still am saying:
"The discrepancy at the top has no statistical significance relative to the rest of the Diamond population."
My argument has been, and always will be, that The >1300 population HAS clearly has a statistically significant distribution in favor of Terran. HOWEVER, this does NOT show correlation to the REST OF DIAMOND.
People are crying OBVIOUS Terran imba. Now, clearly at the super Diamond level, there is statistical support to Terran being overrepresented; HOWEVER, their is no statistical support to the rampant cries of Terran imba form the 300-1000 level Diamond players. That has been my argument and will remain my argument.
CLEARLY THE TOP IS SKEWED. I have yet to say anything otherwise. I have only said that there is NO EVIDENCE to support that this population's distribution is CONSISTENT with the rest of Diamond.
Conclusion: You can cry Terran imba as a 900 player, but you lack evidence to support it from this graph. In fact, this graph suggests, at lower levels, Protoss is more prevalent and overrepresented up to the "super Diamond" level. NOW, if you try to say "IMBA at the top means IMBA everywhere" I can explain why selecting for the most skilled players biases the results when part of the categorical selections MAY impact skill.
Thank you for reading closely how I said that THE TOP OF DIAMOND does not represent DIAMOND, in 4 posts, so you can stop generalizing.
I don't need to insult anyone's classes after all, because your inability to read my actual words proves how poor your education was.
|
I saw some people saying that the reason that Terran was so prevalent in some of the upper leagues was due to many people playing Terran because of the campaign and cries of imba. I hypothesized that if Terran was indeed more popular overall, then it would be shown in a listing of all players' race percentages. If Terran overall was at around 40% and higher than both Protoss and Zerg, then that suggestion might be somewhat founded. I went to sc2ranks.com to check the statistics of all players through all leagues. I will admit, I did not count each individual one by one, but counted 100 people on each page and multiplied the total page count by it, so my raw numbers could be off by up to 99. Here are the results, copypasted from Notepad:
p: 358600 35.9%
z: 193300 19.4%
t: 365200 36.6%
r: 80800 8.1%
+: 997900 100.0%
So, unsurprisingly, Random is at the bottom of the pack by far, Zerg is behind by an incredible 15%, and Terran and Protoss are neck and neck for the lead. This clearly does not support the hypothesis that Terran is more prominent in all leagues due to the campaign and the T IMBA cries. Whether it says anything about the top players individually choosing to play the "better" race is up for debate. No matter what, I would be willing to posit that the game is not balanced at all levels of play, as a huge margin of players feel that Zerg is either underpowered or not fun to play.
If anyone has any problems with my data or my conclusions, please, I'm willing to take criticism.
EDIT: To the above poster, your analysis is slightly incorrect. You assume that players of similar ranking are automatically of similar skill. This is not necessarily true. For example, let's imagine that Zerg was overpowered, so that Zerg players had a significantly easier time beating opposing Protoss and Terran players. A Zerg player of less skill would make it into a higher league of play than they would if Zerg was not overpowered, with a Zerg buildup pooling at the top because there are no higher level players to play against. The lower-level players could certainly complain of imbalance, as the Zerg players who they split about 50-50 with are at a much lower level of skill. This appears to be happening with Terran at the moment, with a buildup of Terran at the top and a more average layout through the rest.
|
On September 03 2010 05:04 Sleight wrote:Show nested quote +On September 03 2010 03:32 Winter_mute wrote:
Imagine you want to examine the color of the ultra rare mexican swamp tree frog (mtf). Now because this frog is soooooo ultra rare you can look at the whole population of MSTFs. 60% are yellow, 30% are blue and 10% are red. You got 360 of these frogs and do a test to see the propability of the colors being randomly distributed. The tests says: No, with a propability of x% these numbers are not random. (see previous post). Congratulations, you found out something!
Now imagine people are coming and telling you that the MSTF only makes up much less than 1% of all the frogs and start talking about random samples blablabla. But you dont care. You said that your results are valid for this one type of frog (or only valid for all people >1200). And your results are pretty clear and significant. And it absolutely does not matter even the slightest how many other frogs there are. NOT EVEN A TINY BIT.
To go back to SC2: even if 200 million billion people were playing zerg in bronze would not mean that the data for the top players (>1200) is wrong. Fact is: there is a disproportional amount of terrans at the top.
Now if you think what would you expect if there was an imbalance: You would expect that the players for race x are shifted by y amount of points and this would be most apparent at the absolut top and the bottom. Hintedi hint hint.
But feel free to do another nice ad hominem and tell me that my university classes sucked and that I also have to study harder. Alternatively you could read about cognitive biases and discover, that people do not modify their views of the world to fit new information but rather fit the information into their views and beliefs. With hillarious results. Great, this is an example of the misunderstanding. Everything posted about the Tree Frog are dead on. Absolutely, if this was a random sampling of a rare population, this could not be more correct. Here is what I said and still am saying: "The discrepancy at the top has no statistical significance relative to the rest of the Diamond population." My argument has been, and always will be, that The >1300 population HAS clearly has a statistically significant distribution in favor of Terran. HOWEVER, this does NOT show correlation to the REST OF DIAMOND. People are crying OBVIOUS Terran imba. Now, clearly at the super Diamond level, there is statistical support to Terran being overrepresented; HOWEVER, their is no statistical support to the rampant cries of Terran imba form the 300-1000 level Diamond players. That has been my argument and will remain my argument. CLEARLY THE TOP IS SKEWED. I have yet to say anything otherwise. I have only said that there is NO EVIDENCE to support that this population's distribution is CONSISTENT with the rest of Diamond. Conclusion: You can cry Terran imba as a 900 player, but you lack evidence to support it from this graph. In fact, this graph suggests, at lower levels, Protoss is more prevalent and overrepresented up to the "super Diamond" level. NOW, if you try to say "IMBA at the top means IMBA everywhere" I can explain why selecting for the most skilled players biases the results when part of the categorical selections MAY impact skill. Thank you for reading closely how I said that THE TOP OF DIAMOND does not represent DIAMOND, in 4 posts, so you can stop generalizing. I don't need to insult anyone's classes after all, because your inability to read my actual words proves how poor your education was.
And again a nice ad hominem about my poor education. Thank you very much. Obviously well educated people take your side, because you are always right. And we all know that people with the "best" education are always right. I mean in science we do not actually discuss stuff and interpretate data, we just ask the guy with the most years as a scientist and his opinion is the truth.
What are you talking about random sampling? There is no random sampling in the frog example. You sample all of them. You cannot generalize your drug testing stuff to everything.
Now lets look at the data again. Ask yourself: would you expect to see a clear imbalance in the diamond population if you consider the match making system? In fact you will not see an imbalance in the rest of the diamond population, because the MMS compensates by matching people with unequal skill: i.e. a skilled player of race a will play an unskilled player of race b. There would have to be a massive, massive imbalance. And consider, that the third race will smooth the imbalances further out.
Just do a graph for yourself. For example you can use a gaussian distribution with 3 different races all at 33% with a small standard deviation. (Maybe even adjust these numbers because you tell yourself, that everone plays terran, because of the campaign.) And now introduce the imbalance. Shift all the players of race1 by 50 or 100 points into 1 direction. Look at what you see. More importantly look in which area you will see the imbalance.
And if you do not have enough statistical evidence, just look at the last tourneys. Look at the last zotac cups, viking cups, Go4SC2 etc. I remeber so many nice cups where there were only terrans in the semi finale. Hell, just watch the ESL stream now. There was 1 (!) single zerg in the best of 8 (dimaga). And he got kicked out in the first round.
However if this is too much work for you, feel free to just question my character or education instead.
|
With Starcraft, it's not so obvious that x race should account for x % of the population at high levels. With such a small sample size at the top of Diamond, the percentages of Terran don't hold as much value, you just need more people at the top before you can make an analysis.
With something simple, such as flipping a coin where there are only two outcomes,it's easy to see when a sample size becomes representative of the whole population. Say there are 1 billion flipped pennies, you should see a probability almost exactly 50% for heads/tails.
Flip 10, and you could easily get 7/10 heads so an inaccurate representation. Flip 100 and It becomes much closer to 50, flip 500 and you are bound to be close to 50% almost every time. So in this case flipping 500 coins is representative of 1 billion coin flips,, less than a thousandth of a percent of the whole population.but this would still hold true for even an infinite amount of coins. (in which case the average would be exactly 50%). Another example is political polls, they contain an extremely small percent of the population but are incredibly accurate.
With Starcraft it's much more complicated, 300 at the top isn't enough to accurately determine if a race is overpowered, All of Diamond yes, but not the top level. Statistics alone won't tell you if a race is overpowered at this point. I'd go on to say that even looking at the entire populations race distribution isn't accurate at the higher levels, because a disproportionate amount of people play Terran as their race because they are used to it. At higher levels of play, many of these people are from Brood War and already have a preferred race because of it's style. We don't know if the reason there are so many ____ is because they are overpowered and it's easier to get there, or because it's a more popular race for higher level players because it fits their style more.
Looking at win percentages isn't enough either, since the match system tries to keep you close to 50%, and will pair you with someone of a higher skill level if you keep winning. In order to know if a race is truly overpowered or not, you would have to rely on individual experiences or knowing the hidden rating in the ladder (which Blizzard obviously knows). If there was a trend emerging where Terran players, on average, were put up against higher level opponents and were still winning, while other races were evenly matched at equal ratings, you could make the conclusion that Terran were overpowered.
Of course this graph is still useful, but it alone will not prove anything.
|
On September 03 2010 05:43 phant wrote:
With Starcraft, it's not so obvious that x race should account for x % of the population at high levels. With such a small sample size at the top of Diamond, the percentages of Terran don't hold as much value, you just need more people at the top before you can make an analysis.
With something simple, such as flipping a coin where there are only two outcomes,it's easy to see when a sample size becomes representative of the whole population. Say there are 1 billion flipped pennies, you should see a probability almost exactly 50% for heads/tails.
Flip 10, and you could easily get 7/10 heads so an inaccurate representation. Flip 100 and It becomes much closer to 50, flip 500 and you are bound to be close to 50% almost every time. So in this case flipping 500 coins is representative of 1 billion coin flips,, less than a thousandth of a percent of the whole population.but this would still hold true for even an infinite amount of coins. (in which case the average would be exactly 50%). Another example is political polls, they contain an extremely small percent of the population but are incredibly accurate.
With Starcraft it's much more complicated, 300 at the top isn't enough to accurately determine if a race is overpowered, All of Diamond yes, but not the top level. Statistics alone won't tell you if a race is overpowered at this point. I'd go on to say that even looking at the entire populations race distribution isn't accurate at the higher levels, because a disproportionate amount of people play Terran as their race because they are used to it. At higher levels of play, many of these people are from Brood War and already have a preferred race because of it's style. We don't know if the reason there are so many ____ is because they are overpowered and it's easier to get there, or because it's a more popular race for higher level players because it fits their style more.
Looking at win percentages isn't enough either, since the match system tries to keep you close to 50%, and will pair you with someone of a higher skill level if you keep winning. In order to know if a race is truly overpowered or not, you would have to rely on individual experiences or knowing the hidden rating in the ladder (which Blizzard obviously knows). If there was a trend emerging where Terran players, on average, were put up against higher level opponents and were still winning, while other races were evenly matched at equal ratings, you could make the conclusion that Terran were overpowered.
Of course this graph is still useful, but it alone will not prove anything.
You sir win this thread.
/endthread
|
Graph obviously shows something, but, as previously stated (sort of), there are so few players at the high end that the data is less helpful about balance. A single player tanking 100 points could throw the graph off by whole percentage points . . . .
Furthermore, the correlation between popularity and balance isn't necessarily perfect. For example, because there are undeniably less Zerg players all around (i.e., if you took silver stats, there would still be less Zerg than Toss/Terran), the stats would be a lot more troubling if there were more Zerg at the top than Terran or Toss.
|
Terran is considered OP in relation to the other 2 races.
Protoss is considered balanced in relation to the other 2 races.
Zerg is considered UP in relation to the other 2 races.
This data correlates with this observation.
What we need now is for a new race to become OP and run the data again.
Hypothetically speaking if this data is related to racial balance and Zerg was the OP race and Terran was the UP race we should see a reverse of the data.
So it is easy to solve the problem.
We intentionally OP Zerg, UP Terran and run the test again. keeping Toss balanced as the control.
We intentionally OP Toss, UP Zerg and run the test again. keeping Terran balanced as the control.
We intentionally OP Terran, UP Toss and run the test again. keeping Zerg balanced as the control.
Run the test over and over until we have sufficient data to conclude that the OP race will end up the most played with the most wins at the highest levels of diamond (1500 +)
It is the simplest way to solve this once and for all.
I'll contact Blizzard and let them know what needs to be done, be right back.
|
It's amazing how some people can remain ignorant or intentionally dismiss clear evidence as faulty when it is actually their own logic that is at fault (since they lack even the most basic understanding of statistics). My guess is that they're trying to get this thread locked by starting a flame war with their awful posting. I hope they get banned before this happens.
I'm just going to quote a few informative posts from earlier in the thread, before it got trolled by certain posters.
On September 02 2010 07:55 Liquid`Tyler wrote:
As far as I know, sc2ranks is pretty damn comprehensive of Diamond, especially high Diamond (where it seems some people have a problem with the "sample size"). Yeah, the number of people in the 1500+ group is small but that doesn't mean that there is a problem with the sample size. These numbers aren't extrapolated from a small population of the 1500+ Diamond group. These numbers directly represent that group.
On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200.
On September 02 2010 08:51 ryanAnger wrote:
Okay, I'm looking at various groupings of the Top Players in the world, and these are the stats (excluding random):
Top 5000 Players in the World - 33.57% (1687) Terran, 25.19% (1266) Zerg, 36.22% (1820) Protoss
Top 2500 - 36.38 (911) T, 23.80 (596) Z, 36.50 (914) P
Top 1000 - 39.80 (398) T, 21.30 (213) Z, 36.90 (369) P
Top 500 - 43.63 (219) T, 20.72 (104) Z, 34.06 (171) P
Top 250 - 51.20 (128) T, 18.80 (47) Z, 28.00 (70) P
Top 100 - 50.50 (51) T, 16.83 (17) Z, 29.70 (30) P
This is essentially the same information the graph shows (supports the same idea) just presented in a manner that the lesser-minded might understand.
Regardless of what the balance issue is at the moment, it is very obvious that Terran is dominating at the highest level, and only an idiot would try to dispute that.
Can we attribute these statistics to overall popularity of the race? Extremely doubtful, considering out of the ~65,000 players in Diamond (Global) Protoss is the most popular with 35%, while Terran is second at 31%. If there were a direct correlation between popularity and overall rank, Protoss should be at the very least tied with Terran at the higher levels, if not leading.
|
Calgary25967 Posts
Teamsolid wrote:
It's amazing how some people can remain ignorant or intentionally dismiss clear evidence as faulty when it is actually their own logic that is at fault (since they lack even the most basic understanding of statistics). My guess is that they're trying to get this thread locked by starting a flame war with their awful posting. I hope they get banned before this happens.
I'm just going to quote a few informative posts from earlier in the thread, before it got trolled by certain posters.
On September 02 2010 07:55 Liquid`Tyler wrote:
As far as I know, sc2ranks is pretty damn comprehensive of Diamond, especially high Diamond (where it seems some people have a problem with the "sample size"). Yeah, the number of people in the 1500+ group is small but that doesn't mean that there is a problem with the sample size. These numbers aren't extrapolated from a small population of the 1500+ Diamond group. These numbers directly represent that group.
On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200.
On September 02 2010 08:51 ryanAnger wrote:
Okay, I'm looking at various groupings of the Top Players in the world, and these are the stats (excluding random):
Top 5000 Players in the World - 33.57% (1687) Terran, 25.19% (1266) Zerg, 36.22% (1820) Protoss
Top 2500 - 36.38 (911) T, 23.80 (596) Z, 36.50 (914) P
Top 1000 - 39.80 (398) T, 21.30 (213) Z, 36.90 (369) P
Top 500 - 43.63 (219) T, 20.72 (104) Z, 34.06 (171) P
Top 250 - 51.20 (128) T, 18.80 (47) Z, 28.00 (70) P
Top 100 - 50.50 (51) T, 16.83 (17) Z, 29.70 (30) P
This is essentially the same information the graph shows (supports the same idea) just presented in a manner that the lesser-minded might understand.
Regardless of what the balance issue is at the moment, it is very obvious that Terran is dominating at the highest level, and only an idiot would try to dispute that.
Can we attribute these statistics to overall popularity of the race? Extremely doubtful, considering out of the ~65,000 players in Diamond (Global) Protoss is the most popular with 35%, while Terran is second at 31%. If there were a direct correlation between popularity and overall rank, Protoss should be at the very least tied with Terran at the higher levels, if not leading.
This is a good post. I removed the embedded quotes (sorry teamsolid) to hope that people read it.
|
okay the better 340 players in the world are terran and the 28k bellow them are protoss, so protoss would be the imba race for the ones that doesnt play pretty godly =P
|
Tbh I think these statistics can't be discarded. They do show a trend, and they are significant in pointing out some de facto problems with the balance...
|
On September 03 2010 06:31 noD wrote:
so protoss would be the imba race for the ones that doesnt play pretty godly =P
that conclusion would be a lot harder to draw. Sucks for Zerg either way
|
On September 02 2010 08:29 Sfydjklm wrote:Show nested quote +On September 02 2010 07:46 cup of joe wrote:
it means absolutely nothing because the sample size is tiny at the right end of the graph thats like saying that oil CEO's are not really wealthy because sample size is too small.
That really made me laugh.
So seriously though... pretty sure there is no question that Terran have too many strong options. Blizzard is going in the right direction with the next patch. These numbers back that up a fair bit, although tournament results are quite important, too.
|
thats like saying that oil CEO's are not really wealthy because sample size is too small.
hmmmm.... it's more like saying "90% of top 10 richest people in the world are male." while those are the actual statistics of the top 10 richest people in the world, are they really that helpful in determining overall balance between men and women?
|
On September 03 2010 07:19 starckr wrote:Show nested quote +thats like saying that oil CEO's are not really wealthy because sample size is too small. hmmmm.... it's more like saying "90% of top 10 richest people in the world are male." while those are the actual statistics of the top 10 richest people in the world, are they really that helpful in determining overall balance between men and women?
Well it would be helpfull under the condition that men and women all earn their money with the same job specific to women or men. And if there were a higher power, that compares individual men and women against each other, rates their performance and then compares them again, adjusts their payment etc.
|
On September 03 2010 06:31 noD wrote:
okay the better 340 players in the world are terran and the 28k bellow them are protoss, so protoss would be the imba race for the ones that doesnt play pretty godly =P
That's... an odd hypothetical scenario, and the only explanation I can think of for it is that there's a very specific skill point that one must pass to be any good with Terran, and once that point is passed, it becomes incredible. Unfortunately, it isn't relevant to this discussion in the slightest as those statistics simply aren't true.
|
This doesn't surprise anyone, I don't think. This data has been available for a long time too.
|
On September 03 2010 06:31 noD wrote:
okay the better 340 players in the world are terran and the 28k bellow them are protoss, so protoss would be the imba race for the ones that doesnt play pretty godly =P
this was pretty much the case in bw: protoss were less mechanically demanding and thus the most popular and most successful race among amateurs and up to the korean semipros. at the same time, protoss were the by far least successful race in terms of starleagues, ie the very top end of the hierarchy.
btw displaying this stuff in percent is kinda misleading as it inflates the graph: if there are more terrans at the cost of fewer zergs at the top end, the distance between the terran bar and the zerg bar increases by twice this change in player base. for a single subsample, e.g. the players above 1.3k, the bars themselves still do make sense as they correspond to the number of players of this population. the comparison between subsamples, e.g. those between 1k and 1.3k and those above 1.3k, will get graphically distorted though. the increase of racial disparity between different regions of diamond ladder seems bigger than it actually is on this figure.
|
On September 03 2010 07:51 Black Gun wrote:Show nested quote +On September 03 2010 06:31 noD wrote:
okay the better 340 players in the world are terran and the 28k bellow them are protoss, so protoss would be the imba race for the ones that doesnt play pretty godly =P this was pretty much the case in bw: protoss were less mechanically demanding and thus the most popular and most successful race among amateurs and up to the korean semipros. at the same time, protoss were the by far least successful race in terms of starleagues, ie the very top end of the hierarchy. btw displaying this stuff in percent is kinda misleading as it inflates the graph: if there are more terrans at the cost of fewer zergs at the top end, the distance between the terran bar and the zerg bar increases by twice this change in player base. for a single subsample, e.g. the players above 1.3k, the bars themselves still do make sense as they correspond to the number of players of this population. the comparison between subsamples, e.g. those between 1k and 1.3k and those above 1.3k, will get graphically distorted though. the increase of racial disparity between different regions of diamond ladder seems bigger than it actually is on this figure.
The number of players at the top can not be discounted, even if it is a smaller amount. If all the SUPERSTAR KOREAN players were all protoss, that would be big deal, even if the amount of players causing that descepancy is small
|
I'll have you know that I have gotten an A in Statistics at my local community college and this looks fine to me. The Highest levels reveal the disparity, but there is no proof that someone who is, say, High Platinum Zerg wouldn't otherwise be a Low-Mid Diamond Terran. I would argue that this is almost certainly the case.
I'm going to go out on a limb and say that, given equal skill- Zerg is going to lose. There is so much more to do, and so much more that can go wrong. I played Terran all though Beta and up to release. Last 4 weeks have been playing Zerg- both at low Diamond level.
I'm not sure if its 100% the "strength of the units" type of imbalance. I may favor the argument that fundamentally, Zerg is just far more difficult to play. The scouting, the macro, the fragility of the units, the need to creep- the need to have More workers and More units than a Terran with Mule- it's just harder mechanically.
Thank you and have a good day.
|
On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200.
This part is actually wrong. The null hypothesis should not be that 1/3 of them prefer Tarran. That would measure how unlikely it is for this deviation to randomly occur IF everyone had an equal chance of choosing any of the races. Even though this model is too broad, it might still work in many cases. However, a much better model would be a Goodness of Fit. The null hypothesis should be that the percent of Tarran players in the high diamond represents the general population of players (all of Bnet). This test will probably show that the high diamond population does not represent the general population (at a given alpha), and thus you can conclude that Tarran players are over represented in the upper rungs of the ladder.
The main reason you would want to do this is because your hypothesis would be affected by many things you would not want to show up in the statistics. What race do more people play? Then people can begin to go, "maybe that's because of the campaign".
However, a Goodness of Fit test takes into account what the actual population is like, and therefore what race is the most played has no bearing as you are just testing whether the subgroup fits the whole.
Now, if you find that the subgroup of the upper rungs of the ladder does not statistically fit the general population, then you can begin to conclude things like "Tarran is OP, which makes them able to move up the ladder easier since the matchmaking system is going to move them up until they have a 50% winrate" or any "Tarran is not OP" reasoning you like.
|
On September 03 2010 08:07 rackdude wrote:Show nested quote +On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200. This part is actually wrong. The null hypothesis should not be that 1/3 of them prefer Tarran. That would measure how unlikely it is for this deviation to randomly occur IF everyone had an equal chance of choosing any of the races. Even though this model is too broad, it might still work in many cases. However, a much better model would be a Goodness of Fit. The null hypothesis should be that the percent of Tarran players in the high diamond represents the general population of players (all of Bnet). This test will probably show that the high diamond population does not represent the general population (at a given alpha), and thus you can conclude that Tarran players are over represented in the upper rungs of the ladder. The main reason you would want to do this is because your hypothesis would be affected by many things you would not want to show up in the statistics. What race do more people play? Then people can begin to go, "maybe that's because of the campaign". However, a Goodness of Fit test takes into account what the actual population is like, and therefore what race is the most played has no bearing as you are just testing whether the subgroup fits the whole. Now, if you find that the subgroup of the upper rungs of the ladder does not statistically fit the general population, then you can begin to conclude things like "Tarran is OP, which makes them able to move up the ladder easier since the matchmaking system is going to move them up until they have a 50% winrate" or any "Tarran is not OP" reasoning you like.
Look at the end of the previous page. I examined this possibility. It's still pretty absurd.
|
+ Show Spoiler +On September 03 2010 06:22 Chill wrote:Teamsolid wrote: It's amazing how some people can remain ignorant or intentionally dismiss clear evidence as faulty when it is actually their own logic that is at fault (since they lack even the most basic understanding of statistics). My guess is that they're trying to get this thread locked by starting a flame war with their awful posting. I hope they get banned before this happens. I'm just going to quote a few informative posts from earlier in the thread, before it got trolled by certain posters. Show nested quote +On September 02 2010 07:55 Liquid`Tyler wrote:
As far as I know, sc2ranks is pretty damn comprehensive of Diamond, especially high Diamond (where it seems some people have a problem with the "sample size"). Yeah, the number of people in the 1500+ group is small but that doesn't mean that there is a problem with the sample size. These numbers aren't extrapolated from a small population of the 1500+ Diamond group. These numbers directly represent that group. Show nested quote +On September 02 2010 08:34 StarDrive wrote:
There are 360 players 1200+. The null hypothesis is that 1/3 of them prefer Terran. We observe around 1/2 of them preferring Terran. Doing some basic statistics with normal approximation of the binomial distribution, the z-score is 6.7. We would observe this Terran favored skew with probability far less than one in a billion. The probability that this Terran favored skew is purely by chance is less than the probability a randomly chosen person has an IQ > 200. Show nested quote +On September 02 2010 08:51 ryanAnger wrote:
Okay, I'm looking at various groupings of the Top Players in the world, and these are the stats (excluding random):
Top 5000 Players in the World - 33.57% (1687) Terran, 25.19% (1266) Zerg, 36.22% (1820) Protoss
Top 2500 - 36.38 (911) T, 23.80 (596) Z, 36.50 (914) P
Top 1000 - 39.80 (398) T, 21.30 (213) Z, 36.90 (369) P
Top 500 - 43.63 (219) T, 20.72 (104) Z, 34.06 (171) P
Top 250 - 51.20 (128) T, 18.80 (47) Z, 28.00 (70) P
Top 100 - 50.50 (51) T, 16.83 (17) Z, 29.70 (30) P
This is essentially the same information the graph shows (supports the same idea) just presented in a manner that the lesser-minded might understand.
Regardless of what the balance issue is at the moment, it is very obvious that Terran is dominating at the highest level, and only an idiot would try to dispute that.
Can we attribute these statistics to overall popularity of the race? Extremely doubtful, considering out of the ~65,000 players in Diamond (Global) Protoss is the most popular with 35%, while Terran is second at 31%. If there were a direct correlation between popularity and overall rank, Protoss should be at the very least tied with Terran at the higher levels, if not leading. This is a good post. I removed the embedded quotes (sorry teamsolid) to hope that people read it.
Hero chill fighting the trolls with appropriate quotes and statistics. Chill fightinggg!
I love how zerg just quickly drops as players get better. Tester has been supporting this evidence since the beginning of beta, anyone who disagrees with what the statistics are showing is just plain ignorant.
|
statistics 101. this is NOT a gaussian distribution, aka BELL curve. the curve should peak in the middle and have a low number of players on the sides, ideally and equal number for each race.
HOWEVER...no one really knows exactly how blizzard's ladder system works. its mind boggling that it can be so skewed. even protoss seems advantaged compared to zerg.
this graph is also skewed. should be # of players not percentages, percentages makes it seem weird.
# of players should also be normalized. this graph is full of fail. author probably has not taken a stats class.
can OP send me the raw data as in...
600 pts level:
xxx # of protoss players
yyy # of zerg ...
...
i'd like to do a better representation.
edit: this can't be a bell curve because its just such a poorly represented graph.
|
One thing I'd like to point out is that the distribution in diamond (or platinum, when that was the highest league in beta) hasn't been constant. Look here: http://www.teamliquid.net/forum/viewmessage.php?topic_id=117123
This is a race distribution of early phase 1 in the beta, when Zerg was much more powerful than its current incarnation due to roaches still being 1 supply and having 2 base armor.
Now, you'd have to make a couple assumptions to disprove the argument that Zerg are unpopular for reasons not related to effectiveness/underpoweredness:
1. The race distribution of all leagues in phase 1 platinum isn't very different from the distribution within phase 1's platinum leagues (i.e., no race was over- or under-represented in platinum at the time).
2. The race distribution of all leagues in the beta would have been similar to the distribution of the live game had the game been released as it was balance-wise on 3/26/2010 (when this data was gathered).
It seems that the non-humanoid insect-like nature of the Zerg didn't bother 29.8% of the platinum players back then. Furthermore, Terran and Protoss are closer to each other and to Zerg (30.1% and 29.2%).
If it were possible to find, it would be interesting to find population data from earlier in the game's development (i.e., back when Zerg was overrepresented and considered "OP" in Korea in phase 1). While it's impossible to say either way, I'd be leaning towards the notion that Zerg population was overall was proportionally higher back then -- and maybe not just on Asian servers.
Furthermore, if one accepts the two assumptions above, one could also say that a race's ease-of-use or viability DOES affect how much of the population chooses to play that race, if you compare this distribution to today's diamond league.
Food for thought.
|
On September 03 2010 09:07 Toxigen wrote:One thing I'd like to point out is that the distribution in diamond (or platinum, when that was the highest league in beta) hasn't been constant. Look here: http://www.teamliquid.net/forum/viewmessage.php?topic_id=117123This is a race distribution of early phase 1 in the beta, when Zerg was much more powerful than its current incarnation due to roaches still being 1 supply and having 2 base armor. Now, you'd have to make a couple assumptions to disprove the argument that Zerg are unpopular for reasons not related to effectiveness/underpoweredness: 1. The race distribution of all leagues in phase 1 platinum isn't very different from the distribution within phase 1's platinum leagues (i.e., no race was over- or under-represented in platinum at the time). 2. The race distribution of all leagues in the beta would have been similar to the distribution of the live game had the game been released as it was balance-wise on 3/26/2010 (when this data was gathered). It seems that the non-humanoid insect-like nature of the Zerg didn't bother 29.8% of the platinum players back then. Furthermore, Terran and Protoss are closer to each other and to Zerg (30.1% and 29.2%). If it were possible to find, it would be interesting to find population data from earlier in the game's development (i.e., back when Zerg was overrepresented and considered "OP" in Korea in phase 1). While it's impossible to say either way, I'd be leaning towards the notion that Zerg population was overall was proportionally higher back then -- and maybe not just on Asian servers. Furthermore, if one accepts the two assumptions above, one could also say that a race's ease-of-use or viability DOES affect how much of the population chooses to play that race, if you compare this distribution to today's diamond league. Food for thought.
this isn't a distribution.
|
On September 03 2010 05:50 starckr wrote:
Graph obviously shows something, but, as previously stated (sort of), there are so few players at the high end that the data is less helpful about balance. A single player tanking 100 points could throw the graph off by whole percentage points . . . .
Furthermore, the correlation between popularity and balance isn't necessarily perfect. For example, because there are undeniably less Zerg players all around (i.e., if you took silver stats, there would still be less Zerg than Toss/Terran), the stats would be a lot more troubling if there were more Zerg at the top than Terran or Toss.
You and everyone that has said this in this thread has no clue what they are talking about.
It shows a trend..and it's not even a vague trend. it's a crystal clear trend that shows that at the highest level of gameplay and really the ONLY place where racial balance comes into play Terrans are dominating.
The end. It's not a sample size, it's a direct representation, know the difference.
|
this graph is also skewed. should be # of players not percentages, percentages makes it seem weird...this graph is full of fail. author probably has not taken a stats class...can OP send me the raw data as in...
@yourwhiteshadow I don't think this is true. If you read my post (aka first post aka OP's post) you will notice that:
+ Show Spoiler +a.) I did not only include the percentage but the player count in each intervall
b.) a graphic which shows what you want to look at which is quite uninteresting due to population's difference
c.) the data is available and is linked in the first post, as also stated several times in this thread and mentioned by various people
d.) the weirdness occures because you want to look at somethign different if you want it to look different do it, everything you need is given to you
e.) the normalization could be taken in account if we figured out how we should do it, this is not clear, so you may think about it (option I. normalize by percentage in each intervall II. normalize over all 3 million sc2 gamers III. normalize as difference from other leagues IV. normalize aka fit all variations to a function and only look at this functions point-density)
f.) we dont have to see a peak in a gauss curve, as it is not needed to look at the middle for certain types of questions since I. we can fit a gauss curve even if we only look at the outliers since it is only basic math we have to use and II. the assumption that we will definetly not see a gauss curve is something possible but unmotivated
g.) it is true that we don't know that the AMM works, this is the reason that I mentioned in my first post, that "Some suggested to compare the racial distribution (RD) at high skill levels as indicator for imbalance or more likely as indicator for the lack of exploitable gamestyle." and therefore the difference on high skill level could be interesting.
h.) that I included the sub 1000pts player does not mean that I am looking at them, I presented a graph with as much information as possible (the reason I excluded the sub 600 pts diamonds is that I am over 600 points and know that I am not skillfull in all aspects of the game)
i.)nternet: BTW did you every published a scientific paper? I doubt it.
k.) If you want to make something better I welcome you, because this is what I want to happen in this thread: People should make assumptions, hypothesis, look at them again, discuss them and see how the patch changes things. This could lead to a better understanding not only about racial preferences but about the communities trends as a whole.
|
On September 03 2010 09:17 Jayme wrote:Show nested quote +On September 03 2010 05:50 starckr wrote:
Graph obviously shows something, but, as previously stated (sort of), there are so few players at the high end that the data is less helpful about balance. A single player tanking 100 points could throw the graph off by whole percentage points . . . .
Furthermore, the correlation between popularity and balance isn't necessarily perfect. For example, because there are undeniably less Zerg players all around (i.e., if you took silver stats, there would still be less Zerg than Toss/Terran), the stats would be a lot more troubling if there were more Zerg at the top than Terran or Toss. You and everyone that has said this in this thread has no clue what they are talking about. It shows a trend..and it's not even a vague trend. it's a crystal clear trend that shows that at the highest level of gameplay and really the ONLY place where racial balance comes into play Terrans are dominating. The end. It's not a sample size, it's a direct representation, know the difference.
Dude you are a friggin retard. Don't call others out when you clearly have no understanding of stats. I mean it's explained statistically on pretty much every page why it's not a trend. Do you really not get that as the population goes down, the error increases significantly?
So the most accurate representations are the earlier ones and they get less and less accurate as you go up. Please tell me you at least understand that much. Because from there it's easy.
Then imagine the point where they are. 35% T, 20% Z, 35% P we'll say at the lowest rating level. The error here can be anything, I'll use plus or minus 1% of what the "real" representation is, because our numbers are just a snapshot in time, not representative of the changes that occur all the time as people play.
Now the next tier up has about half the population of the lowest tier. This means that the error is almost double the error of the previous population, or in this case, around 2%. Then the next tier has half the population of that one. The error here is going to be around 4% and so on and so forth. So when you finally get to your population of 25 at the highest level, the error on what the 'real' numbers are is gigantic.
This is to say, practically, that at the top the people are moving around a lot so the volitility is very high.
The posts saying this includes all players is completely besides the point, that's not what we're comparing. All you can say with respect to that is "there are 11 terran players in the top 25." Saying that because it includes all the numbers it's accurate completely ignores the fact that this is a single snapshot in time. An equally invalid argument would be me saying that in the top 1 player, Zerg is represented 100%, Terran 0, and Protoss 0. Wouldn't that change the look of our "graph" we'll call it?
|
Tray, before you start throwing around fighting words, perhaps you should actually know what you're talking about.
You can't just claim that the results don't show a trend because "the numbers are small". That is not how statistics work. Statisticians use tests to determine significance and confidence. They can actually TELL you how confident you should be in a measurement.
Someone in this thread has already calculated significance for one part of the data. Their result was overwhelmingly in favor of there being a disproportionate amount of terrans in higher level diamond. The odds of it just being a fluke turned out to be one in a BILLION (more or less).
So either you are saying they lied or you are just ignoring the actual statistical reasoning in favor of your cargo cult wishy washy argument that relies more on jargon than mathematical reasoning.
|
Is it just me, or does anyone else think it's laughable that some people think terrans are overrepresented because that's the race you use in the campaign?
I mean, -maybe- in the bronze league there could be slightly more people playing terran, they're true SC beginners and played the campaign before trying out multiplayer. But top level diamond players? Do you REALLY think pro players choose their race because of the single player campaign and not because of which race is the most effective and earns more wins?
Plus there's the fact that we actually have the real numbers. Terran isn't even the most popular race, protoss is. LOL, I guess Artosis must have made a huge impact on people and made them want to choose protoss, right?
I think the fact that the campaign is mostly terran determined the race choice for maybe 1% of the players at best. When asked why there are so many terrans, "because of the campaign" is one of those simple answers that appears to make sense if you don't think about it... but it really isn't true at all.
For the vast majority of people, they don't consider the campaign at all. They try the races out, and stick with the one they like better. And what do people like? Ease of use, fun to use units, variety of units... and most importantly, how powerful the race feels.
Terran is all over those categories. Protoss can be player-friendly too... fewer units to control at once, more powerful individual units, and very strong when in large groups.
Zerg just feels like nothing special... no high damage units, no hard-to-counter units, and the "swarminess" from SC1 is gone. It's pretty easy to see why there are barely half as many zerg players as there are for other races.
|
On September 03 2010 09:27 ReplayArk wrote:Show nested quote +this graph is also skewed. should be # of players not percentages, percentages makes it seem weird...this graph is full of fail. author probably has not taken a stats class...can OP send me the raw data as in... @yourwhiteshadow I don't think this is true. If you read my post (aka first post aka OP's post) you will notice that: + Show Spoiler +a.) I did not only include the percentage but the player count in each intervall
b.) a graphic which shows what you want to look at which is quite uninteresting due to population's difference
c.) the data is available and is linked in the first post, as also stated several times in this thread and mentioned by various people
d.) the weirdness occures because you want to look at somethign different if you want it to look different do it, everything you need is given to you
e.) the normalization could be taken in account if we figured out how we should do it, this is not clear, so you may think about it (option I. normalize by percentage in each intervall II. normalize over all 3 million sc2 gamers III. normalize as difference from other leagues IV. normalize aka fit all variations to a function and only look at this functions point-density)
f.) we dont have to see a peak in a gauss curve, as it is not needed to look at the middle for certain types of questions since I. we can fit a gauss curve even if we only look at the outliers since it is only basic math we have to use and II. the assumption that we will definetly not see a gauss curve is something possible but unmotivated
g.) it is true that we don't know that the AMM works, this is the reason that I mentioned in my first post, that "Some suggested to compare the racial distribution (RD) at high skill levels as indicator for imbalance or more likely as indicator for the lack of exploitable gamestyle." and therefore the difference on high skill level could be interesting.
h.) that I included the sub 1000pts player does not mean that I am looking at them, I presented a graph with as much information as possible (the reason I excluded the sub 600 pts diamonds is that I am over 600 points and know that I am not skillfull in all aspects of the game)
i.)nternet: BTW did you every published a scientific paper? I doubt it.
k.) If you want to make something better I welcome you, because this is what I want to happen in this thread: People should make assumptions, hypothesis, look at them again, discuss them and see how the patch changes things. This could lead to a better understanding not only about racial preferences but about the communities trends as a whole.
/rant
so here is my issue, you can't just look at the tail end of a distribution curve for a few reasons:
1). you have to idea where the average lays
2). because of 1, you no longer know how many std deviation units you are from the avg
3). looking at 1/3 of the picture leaves out a ton of information
we expect to see a bell curve, which is what we would assume would be indicative of a "balanced" game. a balanced game would let 'nature' decide, and ultimately let skill decide which player is better, right? so then why not represent the data as a distribution function of race. most people would expect to see a bell curve if we assume that the game is fair and balanced. however, that is something we don't see, and so we jump to the conclusion, "IMBAAAAA!!!!!!". yet, if we look at the graph again, we clearly see that it isn't a good representation of the data. let's say hypothetically, there are 2k protoss and terran players each, and only 50 zerg players. the way that graph is set up, it would make you think that the game is "IMBA" because the zerg bars are ALWAYS going to be lower than the others...unless magically the top 20 players are zerg and they are the only ones above 1500 (even then, it'd seem weird).
to summarize, what i'm trying to say is that we expect to see a bell curve, but we do not. there is a reason for it. the data is simply NOT normalized. so we really SHOULDN'T be looking for a bell curve. population distribution...well we didn't do it properly...based on my math it looks like everything is well imba, and without getting too technical with errors, deviations, etc, it seems that the races are well represented if you take into account the number of players playing each race.
btw, i play toss, and as much as i hate terran and want to believe the game is imba towards terran, it looks like its not:
https://spreadsheets.google.com/ccc?key=0AhUL1GZVJIr2dHpWVmppNkhjZVNWcW9FQ0E4MkM4dWc&hl=en
|
Tray, ad hominem is generally frowned upon in rational discussion. Calling someone a "friggin retard" does little to support your point to a discerning audience, and tends to shut down the debate. Please stop. You are free to argue your points, just not shout people down and insult them.
To address your actual points: there is no reason to assume that the larger sample size is the accurate one when different portions of the population are under different conditions. For example, let's say we have a town with 10,000 slaves and 10 slaveowners. If we were to measure the average (mean) income in the town, we would get a number between 0 and the income of a slaveowner. Your method is suggesting that we measure the income of the slaves only, because they are the larger sample size. That would get us an incorrect mean (but a correct mode, which is not what we are looking for). In the situation with Terrans in Diamond League, we have a general trend of there being more Terrans the farther up we go. Even before Terran starts dominating the ladder, we still have that trend of more and more Terrans showing up. It is unreasonable to assume that a trend that appears even in the larger sections of the Diamond League must suddenly cut off as soon as we pass the 1200 point mark. You are examining statistics in isolation, which isn't very helpful at all. Here is the core argument that many, many people have repeated throughout this thread: if Terran is overpowered, then Terran players of less skill will rise through the rankings to their respective 50% winrate zones. However, players on the upper end of the spectrum will inevitably "max out" at some point and reach the upper limit of the ranking system. At that point, the Terran players will start to fill up the upper end of the ranking system. This will show itself as a disproportionate representation of Terrans at the upper end of Diamond League, with an increase in Terran players the farther up one goes. That is exactly what this graph shows us. About your "snapshot in time" theory, I'd be glad to examine further if you can produce some more data from earlier since release. However, as things stand, we have a single set of very strong data.
Finally, please don't use ad absurdum. I believe everyone knows that there is a significant difference between one person and twenty, and twenty people and 340. The fact remains that the top Zerg is 50 points ahead of the next closest opponent, a Terran, and has a much lower win/loss ratio than said Terran. It's not that important who ladders more.
|
It's really no surprise that Terran is so high like that. Everyone loves to follow flavors of the month. Sure Terran is pretty powerful right now, but I'm willing to bet that the bigger impact has been everyone complaining about Terran being OP everywhere for the past month or so. Most people on BNet I'm sure want to win, not improve at the game, and so of course they will pick the race that gives them the biggest advantage to win - even if they've only heard about the imbalance but haven't experienced it for themselves.
|
nice stats, didnt know there was so many terrans in top diamond
|
A picture is worth a thousand words. And a graph is worth 10,000. Props to the OP. Hope this gets around to Blizzard.
|
On September 03 2010 09:36 blacktoss wrote:
Tray, before you start throwing around fighting words, perhaps you should actually know what you're talking about.
You can't just claim that the results don't show a trend because "the numbers are small". That is not how statistics work. Statisticians use tests to determine significance and confidence. They can actually TELL you how confident you should be in a measurement.
Someone in this thread has already calculated significance for one part of the data. Their result was overwhelmingly in favor of there being a disproportionate amount of terrans in higher level diamond. The odds of it just being a fluke turned out to be one in a BILLION (more or less).
So either you are saying they lied or you are just ignoring the actual statistical reasoning in favor of your cargo cult wishy washy argument that relies more on jargon than mathematical reasoning.
Yes I can claim that. It's EXACTLY what I'm claiming and it's true. The person you're referring to did the worst math and completely ignored the fact that Terran isn't expected to be 33% representative. It should be relative to the total population of players that play terran if skill is evenly distributed between the three races, which yes, we can assume.
So his one in a billion was not even remotely accurate. In fact these numbers near the tail end are without any doubt are statistically insignificant. He didn't lie, he just didn't know what he was doing.
|
@Savio I think that Blizzard is doing much more in depth analysis of the game balance, they even have the individual player's win rate versus each race and the explicit build order. I think they are in fact running Chi-Square tests on their assumptions, and run algorithm over a set of ten thousand proxy rush games to see when the first Zealots are attacking stuff and if the five second delay will work, before they are explicitly paying tester to look if the balance change will do as they wish.
@yourwhiteshadow Thanks for doing some work, I would like to mention that your normalization is chipped because you watch the player with more than 600 points, why do you do this? There are much more player, if you include the sub 600 points diamonds you will get 64k people. So this is one thing I mentioned in my first respond to you. If you want to normalize - and we surely can do a little stuff with it - we can't simply cut the player off at one end. We should motivate why we do it.
I think you have a good point if you say that to only watch the outliners is bad, but if you got the values: + Show Spoiler +
4,1 8,9E-005
4,2 5,8E-005
4,3 3,8E-005
4,4 2,4E-005
4,5 1,6E-005
4,6 1,0E-005
What I further think is a nice idea you have is the hypothesis that the race's should be fitting a curve over various skill level (if we take league's and points as skill), but I think if we wait and look at the data after the patch we will be able to tell a lot more about some of our assumptions.
|
On September 03 2010 09:57 Acritter wrote:
Tray, ad hominem is generally frowned upon in rational discussion. Calling someone a "friggin retard" does little to support your point to a discerning audience, and tends to shut down the debate. Please stop. You are free to argue your points, just not shout people down and insult them.
To address your actual points: there is no reason to assume that the larger sample size is the accurate one when different portions of the population are under different conditions. For example, let's say we have a town with 10,000 slaves and 10 slaveowners. If we were to measure the average (mean) income in the town, we would get a number between 0 and the income of a slaveowner. Your method is suggesting that we measure the income of the slaves only, because they are the larger sample size. That would get us an incorrect mean (but a correct mode, which is not what we are looking for). In the situation with Terrans in Diamond League, we have a general trend of there being more Terrans the farther up we go. Even before Terran starts dominating the ladder, we still have that trend of more and more Terrans showing up. It is unreasonable to assume that a trend that appears even in the larger sections of the Diamond League must suddenly cut off as soon as we pass the 1200 point mark. You are examining statistics in isolation, which isn't very helpful at all. Here is the core argument that many, many people have repeated throughout this thread: if Terran is overpowered, then Terran players of less skill will rise through the rankings to their respective 50% winrate zones. However, players on the upper end of the spectrum will inevitably "max out" at some point and reach the upper limit of the ranking system. At that point, the Terran players will start to fill up the upper end of the ranking system. This will show itself as a disproportionate representation of Terrans at the upper end of Diamond League, with an increase in Terran players the farther up one goes. That is exactly what this graph shows us. About your "snapshot in time" theory, I'd be glad to examine further if you can produce some more data from earlier since release. However, as things stand, we have a single set of very strong data.
Finally, please don't use ad absurdum. I believe everyone knows that there is a significant difference between one person and twenty, and twenty people and 340. The fact remains that the top Zerg is 50 points ahead of the next closest opponent, a Terran, and has a much lower win/loss ratio than said Terran. It's not that important who ladders more.
Your example is not even remotely close to what is happening here. I explained it correctly. No need to create another analogy. It's explained on every page in this thread. If you're still having a hard time understanding it, google "confidence interval" or "statistical analysis and normal distribution.
The problem people are having is that Race representation should be considered independently and measured against how it changes as the rating goes up. The problem is as the rating goes up, the sample goes down, and error increases exponentially.
Anyway your analogy is completely off. I'm not doing anything remotely close to what you're talking about and my theory about the snapshot is not a theory, it's a statistical fact. Take a look at the spreadsheet above your post there idiot.
|
The problem is as the rating goes up, the sample goes down, and error increases exponentially. @Tray the first and last time I will talk to you, since I like to talk to mannered people and if various claims are brought to you that you might be offensive don't take it as given, but think about it somedays. If you say the error increases exponentially I would like to see the formular you use. What is your method to get the error? If you want you can PM me the latex term and everything I need as reference to understand your point, but still now I didn't get the thing you want to say with this post.
|
On September 03 2010 10:17 Tray wrote:
Yes I can claim that. It's EXACTLY what I'm claiming and it's true. The person you're referring to did the worst math and completely ignored the fact that Terran isn't expected to be 33% representative. It should be relative to the total population of players that play terran if skill is evenly distributed between the three races, which yes, we can assume.
So his one in a billion was not even remotely accurate. In fact these numbers near the tail end are without any doubt are statistically insignificant. He didn't lie, he just didn't know what he was doing.
You are right, he should have used ~40% rather than 33%. But the rest of his analysis was correct. The odds of this distribution being caused by chance are way past explainable by random chance. He provided numbers, while you have merely asserted that the sample numbers are too small. "In fact these numbers near the tail end are without any doubt are statistically insignificant". You have no numbers to support this assertion.
|
On September 03 2010 10:24 ReplayArk wrote:Show nested quote +The problem is as the rating goes up, the sample goes down, and error increases exponentially. @Tray the first and last time I will talk to you, since I like to talk to mannered people and if various claims are brought to you that you might be offensive don't take it as given, but think about it somedays. If you say the error increases exponentially I would like to see the formular you use. What is your method to get the error? If you want you can PM me the latex term and everything I need as reference to understand your point, but still now I didn't get the thing you want to say with this post.
Maybe you should just look it up for yourself and get smarter instead of being like everyone else in here and demanding I teach you all how statistics work.
This is 5 seconds of google. http://en.wikipedia.org/wiki/Sample_size
A typical example would be when a statistician wishes to estimate the arithmetic mean of a quantitative random variable (for example, the height of a person). Assuming that they have a random sample with independent observations, and also that the variability of the population (as measured by the standard deviation σ) is known, then the standard error of the sample mean is given by the formula:
σ/sq.root(n)
As N becomes smaller and smaller, where N is the population of the sample size, the error increases.
if n = 25, error = std/5 = 20% (std = 1 for simplicity, doesn't matter what you use, it's all relative)
if n = 50, error = std/7.07 = 14%
if n = 100, error = std/10 = 10%
if n = 1000, error = std/31.6 = 3%
And I'm going to go on because some of you guys really need the education.
So this means that we can with a 95% certainly that the real number of zerg representation at n= 25 is about 20% (or whatever it is) plus or minus 20%. This means that we're 95% sure the 'real' value is between 0-40%, but 5% of the time it will be outside that range.
This is the correct way to analyze this data including sampling error, which in our case takes into account the variability of player ratings.
Also this is why the normalized numbers are all fairly close relative to total population that plays that race at the n = big numbers, but why it's highly skewed at n = small numbers.
|
On September 03 2010 07:54 cursor wrote:
I'm not sure if its 100% the "strength of the units" type of imbalance. I may favor the argument that fundamentally, Zerg is just far more difficult to play. The scouting, the macro, the fragility of the units, the need to creep- the need to have More workers and More units than a Terran with Mule- it's just harder mechanically.
Don't forget the Terran's dominant versatility in terms of weapon ranges and splash damage compared to the other races. 
|
Yes I can claim that. It's EXACTLY what I'm claiming and it's true. The person you're referring to did the worst math and completely ignored the fact that Terran isn't expected to be 33% representative. It should be relative to the total population of players that play terran if skill is evenly distributed between the three races, which yes, we can assume.
So his one in a billion was not even remotely accurate. In fact these numbers near the tail end are without any doubt are statistically insignificant. He didn't lie, he just didn't know what he was doing.
Here's the kicker genius: You're fighting a straw man. The claim made was that Terran make up a disproportionate number of high level diamond players such that the deviation from an even distribution is not due to random chance.
The claim is CORRECT. The math is CORRECT. You are the one coming in here and putting words in peoples' mouths and then saying "haha no u r rong I will now insult you".
I don't think you know how statistics works at all. The hypothesis tested was the hypothesis that Terran should make up 33% of the racial distribution in high diamond league (I am not sure if he took into account Random). The statistical test he used showed with high confidence that this hypothesis was false. End of story that is all the reasoning used.
You talk about statistical insignificance but your criticism does not address the point's validity, it attacks it on the basis that "it is not the right hypothesis". Sorry, but that does not say anything about the conclusion drawn.
Maybe you should just look it up for yourself and get smarter instead of being like everyone else in here and demanding I teach you all how statistics work.
This is 5 seconds of google.
A typical example would be when a statistician wishes to estimate the arithmetic mean of a quantitative random variable (for example, the height of a person). Assuming that they have a random sample with independent observations, and also that the variability of the population (as measured by the standard deviation σ) is known, then the standard error of the sample mean is given by the formula:
σ/sq.root(n)
As N becomes smaller and smaller, where N is the population of the sample size, the error increases.
if n = 25, error = std/5 (std = 1) = 20%
if n = 50, error = std/7.07 = 14%
if n = 100, error = std/10 = 10%
if n = 1000, error = std/31.6 = 3%
Once again, you are full of shit. You come in here saying you are better than anyone else, throw around a few big words, and then cite google and say "hurdy hur".
It doesn't matter if the variance is high, because when YOU DO THE ACTUAL TEST (instead of bullshitting with jargon), you find that p is very high. So high in fact that you MUST discard the hypothesis that "Terran is not overrepresented in high diamond". You can say "WELL SIGMA IS HIGH" but it doesn't matter. Statistical tests are robust. They take sigma into account.
I am beginning to feel like you don't know anything about statistics at all.
|
On September 03 2010 10:40 blacktoss wrote:Show nested quote +Yes I can claim that. It's EXACTLY what I'm claiming and it's true. The person you're referring to did the worst math and completely ignored the fact that Terran isn't expected to be 33% representative. It should be relative to the total population of players that play terran if skill is evenly distributed between the three races, which yes, we can assume.
So his one in a billion was not even remotely accurate. In fact these numbers near the tail end are without any doubt are statistically insignificant. He didn't lie, he just didn't know what he was doing. Here's the kicker genius: You're fighting a straw man. The claim made was that Terran make up a disproportionate number of high level diamond players such that the deviation from an even distribution is not due to random chance. The claim is CORRECT. The math is CORRECT. You are the one coming in here and putting words in peoples' mouths and then saying "haha no u r rong I will now insult you". I don't think you know how statistics works at all. The hypothesis tested was the hypothesis that Terran should make up 33% of the racial distribution in high diamond league (I am not sure if he took into account Random). The statistical test he used showed with high confidence that this hypothesis was false. End of story that is all the reasoning used. You talk about statistical insignificance but your criticism does not address the point's validity, it attacks it on the basis that "it is not the right hypothesis". Sorry, but that does not say anything about the conclusion drawn. Show nested quote +Maybe you should just look it up for yourself and get smarter instead of being like everyone else in here and demanding I teach you all how statistics work.
This is 5 seconds of google.
A typical example would be when a statistician wishes to estimate the arithmetic mean of a quantitative random variable (for example, the height of a person). Assuming that they have a random sample with independent observations, and also that the variability of the population (as measured by the standard deviation σ) is known, then the standard error of the sample mean is given by the formula:
σ/sq.root(n)
As N becomes smaller and smaller, where N is the population of the sample size, the error increases.
if n = 25, error = std/5 (std = 1) = 20%
if n = 50, error = std/7.07 = 14%
if n = 100, error = std/10 = 10%
if n = 1000, error = std/31.6 = 3% Once again, you are full of shit. You come in here saying you are better than anyone else, throw around a few big words, and then cite google and say "hurdy hur". It doesn't matter if the variance is high, because when YOU DO THE ACTUAL TEST (instead of bullshitting with jargon), you find that p is very high. So high in fact that you MUST discard the hypothesis that "Terran is not overrepresented in high diamond". You can say "WELL SIGMA IS HIGH" but it doesn't matter. Statistical tests are robust. They take sigma into account. I am beginning to feel like you don't know anything about statistics at all.
Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable.
As the sample drops, the error in the 'real' value increases and becomes less accurate. Period.
Stay in school.
|
@Tray n^(-1/2) is not exponential while e^(n) is.
|
On September 03 2010 10:44 Tray wrote:Show nested quote +On September 03 2010 10:40 blacktoss wrote:Yes I can claim that. It's EXACTLY what I'm claiming and it's true. The person you're referring to did the worst math and completely ignored the fact that Terran isn't expected to be 33% representative. It should be relative to the total population of players that play terran if skill is evenly distributed between the three races, which yes, we can assume.
So his one in a billion was not even remotely accurate. In fact these numbers near the tail end are without any doubt are statistically insignificant. He didn't lie, he just didn't know what he was doing. Here's the kicker genius: You're fighting a straw man. The claim made was that Terran make up a disproportionate number of high level diamond players such that the deviation from an even distribution is not due to random chance. The claim is CORRECT. The math is CORRECT. You are the one coming in here and putting words in peoples' mouths and then saying "haha no u r rong I will now insult you". I don't think you know how statistics works at all. The hypothesis tested was the hypothesis that Terran should make up 33% of the racial distribution in high diamond league (I am not sure if he took into account Random). The statistical test he used showed with high confidence that this hypothesis was false. End of story that is all the reasoning used. You talk about statistical insignificance but your criticism does not address the point's validity, it attacks it on the basis that "it is not the right hypothesis". Sorry, but that does not say anything about the conclusion drawn. Maybe you should just look it up for yourself and get smarter instead of being like everyone else in here and demanding I teach you all how statistics work.
This is 5 seconds of google.
A typical example would be when a statistician wishes to estimate the arithmetic mean of a quantitative random variable (for example, the height of a person). Assuming that they have a random sample with independent observations, and also that the variability of the population (as measured by the standard deviation σ) is known, then the standard error of the sample mean is given by the formula:
σ/sq.root(n)
As N becomes smaller and smaller, where N is the population of the sample size, the error increases.
if n = 25, error = std/5 (std = 1) = 20%
if n = 50, error = std/7.07 = 14%
if n = 100, error = std/10 = 10%
if n = 1000, error = std/31.6 = 3% Once again, you are full of shit. You come in here saying you are better than anyone else, throw around a few big words, and then cite google and say "hurdy hur". It doesn't matter if the variance is high, because when YOU DO THE ACTUAL TEST (instead of bullshitting with jargon), you find that p is very high. So high in fact that you MUST discard the hypothesis that "Terran is not overrepresented in high diamond". You can say "WELL SIGMA IS HIGH" but it doesn't matter. Statistical tests are robust. They take sigma into account. I am beginning to feel like you don't know anything about statistics at all. Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable. As the sample drops, the error in the 'real' value increases and becomes less accurate. Period. Stay in school.
The game isn't supposed to balanced around a 10000 sample size its supposed to be balanced around the top top top tier which may be 250 people or less. the currents stats say Terran is most represented in top tier, meaning that in the given time its easier to learn, play, and be pro as terran, which is an overpowered quality.
Even if there were 20x more Terran than Zerg, and the game was balanced. That is not good and changes to balance should be made so representation is even.
|
On September 03 2010 10:47 ReplayArk wrote:
@Tray n^(-1/2) is not exponential while e^(n) is.
You're right it's not exponential. It's 1/sq root of n, as I showed 2 posts ago. Poor recollection on my part, I apologize.
|
Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable.
As the sample drops, the error in the 'real' value increases and becomes less accurate. Period.
Stay in school.
Ok, genius. So you're disputing the hypothesis instead of the statistics.
So don't try to argue statistics when your point has nothing to do with the statistics. You think the conclusion does not yield any information about the distribution of players. That has nothing to do with the methodology used or the mathematics used.
And yet you act like a complete jackass and argue the statistics. You say "no they are wrong". No, they are right. Your point is completely orthogonal to that. So I'm going to just assume you are the worst communicator in the world, especially since you have the manners of a seventh grader.
Also, just because the confidence intervals are wider and your p value is going to smaller than if sigma were small, you can still test the hypothesis and draw conclusions.
So stop being a complete asshole, you know a lot of fancy words, but your understanding of them is near nil. You don't know how to interpret values like sigma or p or r or a confidence interval. You know what they are but you have no idea how to interpret them.
So in every single way, you come off as a prick who doesn't know how to put together his own argument.
ANYWAY
Go ahead and find the total proportion of Terran in the entire SC2 population and then test the hypothesis that 1300+ diamond players follow that proportion. Go ahead, show us your gosu statistical acumen. Or are you going to dodge and continue to strut around like a cock?
|
On September 03 2010 10:53 pileopoop wrote:Show nested quote +On September 03 2010 10:44 Tray wrote:On September 03 2010 10:40 blacktoss wrote:Yes I can claim that. It's EXACTLY what I'm claiming and it's true. The person you're referring to did the worst math and completely ignored the fact that Terran isn't expected to be 33% representative. It should be relative to the total population of players that play terran if skill is evenly distributed between the three races, which yes, we can assume.
So his one in a billion was not even remotely accurate. In fact these numbers near the tail end are without any doubt are statistically insignificant. He didn't lie, he just didn't know what he was doing. Here's the kicker genius: You're fighting a straw man. The claim made was that Terran make up a disproportionate number of high level diamond players such that the deviation from an even distribution is not due to random chance. The claim is CORRECT. The math is CORRECT. You are the one coming in here and putting words in peoples' mouths and then saying "haha no u r rong I will now insult you". I don't think you know how statistics works at all. The hypothesis tested was the hypothesis that Terran should make up 33% of the racial distribution in high diamond league (I am not sure if he took into account Random). The statistical test he used showed with high confidence that this hypothesis was false. End of story that is all the reasoning used. You talk about statistical insignificance but your criticism does not address the point's validity, it attacks it on the basis that "it is not the right hypothesis". Sorry, but that does not say anything about the conclusion drawn. Maybe you should just look it up for yourself and get smarter instead of being like everyone else in here and demanding I teach you all how statistics work.
This is 5 seconds of google.
A typical example would be when a statistician wishes to estimate the arithmetic mean of a quantitative random variable (for example, the height of a person). Assuming that they have a random sample with independent observations, and also that the variability of the population (as measured by the standard deviation σ) is known, then the standard error of the sample mean is given by the formula:
σ/sq.root(n)
As N becomes smaller and smaller, where N is the population of the sample size, the error increases.
if n = 25, error = std/5 (std = 1) = 20%
if n = 50, error = std/7.07 = 14%
if n = 100, error = std/10 = 10%
if n = 1000, error = std/31.6 = 3% Once again, you are full of shit. You come in here saying you are better than anyone else, throw around a few big words, and then cite google and say "hurdy hur". It doesn't matter if the variance is high, because when YOU DO THE ACTUAL TEST (instead of bullshitting with jargon), you find that p is very high. So high in fact that you MUST discard the hypothesis that "Terran is not overrepresented in high diamond". You can say "WELL SIGMA IS HIGH" but it doesn't matter. Statistical tests are robust. They take sigma into account. I am beginning to feel like you don't know anything about statistics at all. Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable. As the sample drops, the error in the 'real' value increases and becomes less accurate. Period. Stay in school. The game isn't supposed to balanced around a 10000 sample size its supposed to be balanced around the top top top tier which may be 250 people or less. the currents stats say Terran is most represented in top tier, meaning that in the given time its easier to learn, play, and be pro as terran, which is an overpowered quality.
This is not a correct statement. Blizzard is not balancing the game around statistically insignificant data.
Likely they are using data of a majority of Diamond and then looking at individual games from tournaments and very top ranked players to try to manually spot what we refer to as 'cheese' and 'abuse.' This is because the game is evolving and probably does so from the top down. So you could make an argument that the top players have figured out how to abuse terran, but you would need more data than we have her to prove it. As I just showed, the error in the higher level samples makes that data incredbly volitile.
|
On September 03 2010 10:56 Tray wrote:Show nested quote +On September 03 2010 10:53 pileopoop wrote:On September 03 2010 10:44 Tray wrote:On September 03 2010 10:40 blacktoss wrote:Yes I can claim that. It's EXACTLY what I'm claiming and it's true. The person you're referring to did the worst math and completely ignored the fact that Terran isn't expected to be 33% representative. It should be relative to the total population of players that play terran if skill is evenly distributed between the three races, which yes, we can assume.
So his one in a billion was not even remotely accurate. In fact these numbers near the tail end are without any doubt are statistically insignificant. He didn't lie, he just didn't know what he was doing. Here's the kicker genius: You're fighting a straw man. The claim made was that Terran make up a disproportionate number of high level diamond players such that the deviation from an even distribution is not due to random chance. The claim is CORRECT. The math is CORRECT. You are the one coming in here and putting words in peoples' mouths and then saying "haha no u r rong I will now insult you". I don't think you know how statistics works at all. The hypothesis tested was the hypothesis that Terran should make up 33% of the racial distribution in high diamond league (I am not sure if he took into account Random). The statistical test he used showed with high confidence that this hypothesis was false. End of story that is all the reasoning used. You talk about statistical insignificance but your criticism does not address the point's validity, it attacks it on the basis that "it is not the right hypothesis". Sorry, but that does not say anything about the conclusion drawn. Maybe you should just look it up for yourself and get smarter instead of being like everyone else in here and demanding I teach you all how statistics work.
This is 5 seconds of google.
A typical example would be when a statistician wishes to estimate the arithmetic mean of a quantitative random variable (for example, the height of a person). Assuming that they have a random sample with independent observations, and also that the variability of the population (as measured by the standard deviation σ) is known, then the standard error of the sample mean is given by the formula:
σ/sq.root(n)
As N becomes smaller and smaller, where N is the population of the sample size, the error increases.
if n = 25, error = std/5 (std = 1) = 20%
if n = 50, error = std/7.07 = 14%
if n = 100, error = std/10 = 10%
if n = 1000, error = std/31.6 = 3% Once again, you are full of shit. You come in here saying you are better than anyone else, throw around a few big words, and then cite google and say "hurdy hur". It doesn't matter if the variance is high, because when YOU DO THE ACTUAL TEST (instead of bullshitting with jargon), you find that p is very high. So high in fact that you MUST discard the hypothesis that "Terran is not overrepresented in high diamond". You can say "WELL SIGMA IS HIGH" but it doesn't matter. Statistical tests are robust. They take sigma into account. I am beginning to feel like you don't know anything about statistics at all. Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable. As the sample drops, the error in the 'real' value increases and becomes less accurate. Period. Stay in school. The game isn't supposed to balanced around a 10000 sample size its supposed to be balanced around the top top top tier which may be 250 people or less. the currents stats say Terran is most represented in top tier, meaning that in the given time its easier to learn, play, and be pro as terran, which is an overpowered quality. This is not a correct statement. Blizzard is not balancing the game around statistically insignificant data. Likely they are using data of a majority of Diamond and then looking at individual games from tournaments and very top ranked players to try to manually spot what we refer to as 'cheese' and 'abuse.' This is because the game is evolving and probably does so from the top down. So you could make an argument that the top players have figured out how to abuse terran, but you would need more data than we have her to prove it. As I just showed, the error in the higher level samples makes that data incredbly volitile.
Making a balance changed with current data isn't going to ruin the game. At worst Terran will start getting rolled and you can balance again.
|
These statistics are going to be, as always, the effect of a number of factors.
Go look at B level + on ICCUP. It's not going to be even distribution, there's going to be significantly higher number of P players. This ratio probably skews going up to A level.
What do any of those statistics say? Nothing.
--
Not to mention the ladder is a wildly inaccurate representation of actual skill. Points can be whored by massgame, and possibly many top players don't even play the ladder.
|
On September 03 2010 10:55 blacktoss wrote:Show nested quote +Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable.
As the sample drops, the error in the 'real' value increases and becomes less accurate. Period.
Stay in school. Ok, genius. So you're disputing the hypothesis instead of the statistics. So don't try to argue statistics when your point has nothing to do with the statistics. You think the conclusion does not yield any information about the distribution of players. That has nothing to do with the methodology used or the mathematics used. And yet you act like a complete jackass and argue the statistics. You say "no they are wrong". No, they are right. Your point is completely orthogonal to that. So I'm going to just assume you are the worst communicator in the world, especially since you have the manners of a seventh grader. Also, just because the confidence intervals are wider and your p value is going to smaller than if sigma were small, you can still test the hypothesis and draw conclusions. So stop being a complete asshole, you know a lot of fancy words, but your understanding of them is near nil. You don't know how to interpret values like sigma or p or r or a confidence interval. You know what they are but you have no idea how to interpret them. So in every single way, you come off as a prick who doesn't know how to put together his own argument. ANYWAY Go ahead and find the total proportion of Terran in the entire SC2 population and then test the hypothesis that 1300+ diamond players follow that proportion. Go ahead, show us your gosu statistical acumen. Or are you going to dodge and continue to strut around like a cock?
I'm going to go to sleep. You can do it yourself, it's not too hard. Take the spreadsheet from the previous page that someone was kind enough to post with the nominal numbers. These numbers take into account the total race distribution relative to the % of people that play that race. This however does not include the error.
To find the error simply take the equation I just gave, 1/ sqroot(n) to find the error in those numbers. Now I forget from here if you would have to divide it by 3, or do a more complicated statistical analysis to figure out how much error applies to each normalized race %, but I'm sure if you read the wikipedia page on sampling you can figure out the right way to account for it.
It doesn't matter though because it's obvious from how much the error increases as the sample drops that the 'real' numbers will vary so much at the low population it'll look something like this (with 95% confidence interval); T = 10-50%, Z = 5-35%, P = 10-50%. Very obviously not reliable data because of the huge volitility. More complicated analysis can be done, but it's not necessary.
|
Sooo why is this still up when the data is 100 % wrong and he ethier really really sucks at making graphs or did it on purpose.
Ethier way why has this not been closed lol people keep coming to this thinking its real data they can belive in when its all just crap. Like 60 % of these posts are all saying wow that graph sure tells it like it is or what ever beliveing it.
Go to http://sc2ranks.com/stats/league/all/1/all
For the real #'s and not data that is manipulated or false.
|
On September 03 2010 10:59 pileopoop wrote:Show nested quote +On September 03 2010 10:56 Tray wrote:On September 03 2010 10:53 pileopoop wrote:On September 03 2010 10:44 Tray wrote:On September 03 2010 10:40 blacktoss wrote:Yes I can claim that. It's EXACTLY what I'm claiming and it's true. The person you're referring to did the worst math and completely ignored the fact that Terran isn't expected to be 33% representative. It should be relative to the total population of players that play terran if skill is evenly distributed between the three races, which yes, we can assume.
So his one in a billion was not even remotely accurate. In fact these numbers near the tail end are without any doubt are statistically insignificant. He didn't lie, he just didn't know what he was doing. Here's the kicker genius: You're fighting a straw man. The claim made was that Terran make up a disproportionate number of high level diamond players such that the deviation from an even distribution is not due to random chance. The claim is CORRECT. The math is CORRECT. You are the one coming in here and putting words in peoples' mouths and then saying "haha no u r rong I will now insult you". I don't think you know how statistics works at all. The hypothesis tested was the hypothesis that Terran should make up 33% of the racial distribution in high diamond league (I am not sure if he took into account Random). The statistical test he used showed with high confidence that this hypothesis was false. End of story that is all the reasoning used. You talk about statistical insignificance but your criticism does not address the point's validity, it attacks it on the basis that "it is not the right hypothesis". Sorry, but that does not say anything about the conclusion drawn. Maybe you should just look it up for yourself and get smarter instead of being like everyone else in here and demanding I teach you all how statistics work.
This is 5 seconds of google.
A typical example would be when a statistician wishes to estimate the arithmetic mean of a quantitative random variable (for example, the height of a person). Assuming that they have a random sample with independent observations, and also that the variability of the population (as measured by the standard deviation σ) is known, then the standard error of the sample mean is given by the formula:
σ/sq.root(n)
As N becomes smaller and smaller, where N is the population of the sample size, the error increases.
if n = 25, error = std/5 (std = 1) = 20%
if n = 50, error = std/7.07 = 14%
if n = 100, error = std/10 = 10%
if n = 1000, error = std/31.6 = 3% Once again, you are full of shit. You come in here saying you are better than anyone else, throw around a few big words, and then cite google and say "hurdy hur". It doesn't matter if the variance is high, because when YOU DO THE ACTUAL TEST (instead of bullshitting with jargon), you find that p is very high. So high in fact that you MUST discard the hypothesis that "Terran is not overrepresented in high diamond". You can say "WELL SIGMA IS HIGH" but it doesn't matter. Statistical tests are robust. They take sigma into account. I am beginning to feel like you don't know anything about statistics at all. Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable. As the sample drops, the error in the 'real' value increases and becomes less accurate. Period. Stay in school. The game isn't supposed to balanced around a 10000 sample size its supposed to be balanced around the top top top tier which may be 250 people or less. the currents stats say Terran is most represented in top tier, meaning that in the given time its easier to learn, play, and be pro as terran, which is an overpowered quality. This is not a correct statement. Blizzard is not balancing the game around statistically insignificant data. Likely they are using data of a majority of Diamond and then looking at individual games from tournaments and very top ranked players to try to manually spot what we refer to as 'cheese' and 'abuse.' This is because the game is evolving and probably does so from the top down. So you could make an argument that the top players have figured out how to abuse terran, but you would need more data than we have her to prove it. As I just showed, the error in the higher level samples makes that data incredbly volitile. Making a balance changed with current data isn't going to ruin the game. At worst Terran will start getting rolled and you can balance again.
The problem is that the data doesn't show that Terran is overrepresented within a 95% confidence at the highest level. That's why they balance around the numbers that are statistically significant, which based on the diamond players 600+, are almost perfectly even. (see previous page google spreadsheet link)
|
On September 03 2010 10:44 Tray wrote:
Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable.
As the sample drops, the error in the 'real' value increases and becomes less accurate. Period.
Stay in school.
Tray, I'm going to show you are wrong beyond a reasonable doubt. I will use the calculator found here: http://stattrek.com/Tables/Binomial.aspx
This calculator does not rely on binomial distributions. It uses discrete calculations, just as the ladder behaves.
From the spreadsheet linked earlier, there are 27862 players with 600 or more points in diamond. Terran makes up 30.6% of that. Thus we can treat the success of a Terran player as 30.6%, or .306%. In other words, if player skill is equally distribute (and you said yourself that we should assume this), 30.6% of the top 25 players should be Terran.
So we plug .306 as the "Probability of success on a single trial", because this is the fraction of Terran players. Number of trials is 20, because that is the edge of the ladder ranking we are looking at. There are 10 Terran players in the top 20 as of right now (sc2ranks.com). This is the number of successes.
Now press the calculate button. We are interested in Cumulative Probability: P(X >= 10). It will spit out 0.0544, or about 5%. In other words, there is a ~5% chance for 10 of the top 20 to be Terran. In other other words, there is a 95% probably that the distribution of Terrans in the top 20 is not due to chance. This is good enough for scientific publication.
Sample size has nothing to do with it. Plug in a fake coin flip. Let's say our number events is 2, and we have 2 hits. Probability of success is 0.5. The odds of getting 2 (or more, but that's irrelevent here) is 0.25. Thus I could say with 75% certainty that the coin was rigged. I will be right 75% of the time and wrong 25% of the time. That's how it works.
We're talking a lot more than 75% certainty. 95% certainty. The distribution of players falls outside what would be expected by chance, and sample size has nothing to do with it. (Ignoring the fact that this isn't technically a sample but is rather the complete reality.)
There are only three reasonable explanations:
1. We are in the 5% zone, and this Terran distribution is really just a fluke.
2. Terran is overpowered.
3. Skilled players tend to play Terran more than the other races, and skill distribution is not equal (which was one of our key assumptions).
This is math and it is fact. The Terran distribution falls outside random chance. I am 95% sure of this statement, which is a lot better than you can provide.
|
On September 03 2010 05:16 Acritter wrote:
t: 365200 36.6%
This is the percentage of Terran players in all of Starcraft 2. It is not even close to being accurately represented in upper Diamond League. Tray, please stop telling us all to compare this to data that has already been obtained and which contradicts your points. The fact remains that the high proportion of Terran players in upper Diamond cannot be dismissed as an outlier.
|
On September 03 2010 11:06 Sentient wrote:Show nested quote +On September 03 2010 10:44 Tray wrote:
Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable.
As the sample drops, the error in the 'real' value increases and becomes less accurate. Period.
Stay in school. Tray, I'm going to show you are wrong beyond a reasonable doubt. I will use the calculator found here: http://stattrek.com/Tables/Binomial.aspxThis calculator does not rely on binomial distributions. It uses discrete calculations, just as the ladder behaves. From the spreadsheet linked earlier, there are 27862 players with 600 or more points in diamond. Terran makes up 30.6% of that. Thus we can treat the success of a Terran player as 30.6%, or .306%. In other words, if player skill is equally distribute (and you said yourself that we should assume this), 30.6% of the top 25 players should be Terran. So we plug .306 as the "Probability of success on a single trial", because this is the fraction of Terran players. Number of trials is 20, because that is the edge of the ladder ranking we are looking at. There are 10 Terran players in the top 20 as of right now (sc2ranks.com). This is the number of successes. Now press the calculate button. We are interested in Cumulative Probability: P(X >= 10). It will spit out 0.0544, or about 5%. In other words, there is a ~5% chance for 10 of the top 20 to be Terran. In other other words, there is a 95% probably that the distribution of Terrans in the top 20 is not due to chance. This is good enough for scientific publication. Sample size has nothing to do with it. Plug in a fake coin flip. Let's say our number events is 2, and we have 2 hits. Probability of success is 0.5. The odds of getting 2 (or more, but that's irrelevent here) is 0.25. Thus I could say with 75% certainty that the dice were rigged. I will be right 75% of the time and wrong 25% of the time. That's how it works. We're talking a lot more than 75% certainty. 95% certainty. The distribution of players falls outside what would be expected by chance, and sample size has nothing to do with it. (Ignoring the fact that this isn't technically a sample but is rather the complete reality.) There are only three reasonable explanations: 1. We are in the 5% zone, and this Terran distribution is really just a fluke. 2. Terran is overpowered. 3. Skilled players tend to play Terran more than the other races, and skill distribution is not equal (which was one of our key assumptions). This is math and it is fact. The Terran distribution falls outside random chance. I am 95% sure of this statement, which is a lot better than you can provide.
You're right, this is math. But this is not a correct analysis. You're actually arguing something completely different. I know you're not very smart so you don't probably know that. You probably think you compared the same thing as me.
You're taking the expected number of terrans per player to be .3. Thus if there are 20 people you're saying that 6 of them are expected to be Terran. The actual number is 10 in this case so you're saying that since that's 4 points away from 6, it only has a 5% chance to be true.
Now lets take the real number of % of players who play Terran. 40%. Using your same analysis we would expect out of the top 20 people, 8 of them should be Terran. Now do your cumulative probability and tell me that it's still outside the norm.
There I just used your own analysis to prove you wrong. Feel free to do me now with mine from before.
User was temp banned for this post.
|
On September 03 2010 11:07 Acritter wrote:This is the percentage of Terran players in all of Starcraft 2. It is not even close to being accurately represented in upper Diamond League. Tray, please stop telling us all to compare this to data that has already been obtained and which contradicts your points. The fact remains that the high proportion of Terran players in upper Diamond cannot be dismissed as an outlier.
The 'fact' is that it is dismissed as an outlier. Without any debate. Those numbers are not relevant. Use your value in the guy above me's calculation. You will see that you cannot say with a 95% certainty that they are overrepresented. Period.
Goodbye idiots. I hope they lock this thread because people are only getting dumber coming here.
|
On September 03 2010 10:40 blacktoss wrote:Show nested quote +Yes I can claim that. It's EXACTLY what I'm claiming and it's true. The person you're referring to did the worst math and completely ignored the fact that Terran isn't expected to be 33% representative. It should be relative to the total population of players that play terran if skill is evenly distributed between the three races, which yes, we can assume.
So his one in a billion was not even remotely accurate. In fact these numbers near the tail end are without any doubt are statistically insignificant. He didn't lie, he just didn't know what he was doing. Here's the kicker genius: You're fighting a straw man. The claim made was that Terran make up a disproportionate number of high level diamond players such that the deviation from an even distribution is not due to random chance. The claim is CORRECT. The math is CORRECT. You are the one coming in here and putting words in peoples' mouths and then saying "haha no u r rong I will now insult you". I don't think you know how statistics works at all. The hypothesis tested was the hypothesis that Terran should make up 33% of the racial distribution in high diamond league (I am not sure if he took into account Random). The statistical test he used showed with high confidence that this hypothesis was false. End of story that is all the reasoning used. You talk about statistical insignificance but your criticism does not address the point's validity, it attacks it on the basis that "it is not the right hypothesis". Sorry, but that does not say anything about the conclusion drawn. Show nested quote +Maybe you should just look it up for yourself and get smarter instead of being like everyone else in here and demanding I teach you all how statistics work.
This is 5 seconds of google.
A typical example would be when a statistician wishes to estimate the arithmetic mean of a quantitative random variable (for example, the height of a person). Assuming that they have a random sample with independent observations, and also that the variability of the population (as measured by the standard deviation σ) is known, then the standard error of the sample mean is given by the formula:
σ/sq.root(n)
As N becomes smaller and smaller, where N is the population of the sample size, the error increases.
if n = 25, error = std/5 (std = 1) = 20%
if n = 50, error = std/7.07 = 14%
if n = 100, error = std/10 = 10%
if n = 1000, error = std/31.6 = 3% Once again, you are full of shit. You come in here saying you are better than anyone else, throw around a few big words, and then cite google and say "hurdy hur". It doesn't matter if the variance is high, because when YOU DO THE ACTUAL TEST (instead of bullshitting with jargon), you find that p is very high. So high in fact that you MUST discard the hypothesis that "Terran is not overrepresented in high diamond". You can say "WELL SIGMA IS HIGH" but it doesn't matter. Statistical tests are robust. They take sigma into account. I am beginning to feel like you don't know anything about statistics at all.
He is correct that the wrong test was ran. It shouldn't be a z-test but rather a Chi Square or Goodness of Fit Test since that would take into account the % Tarran in the whole population as the expected value. However, it doesn't really matter since the graph is so clear that it's pretty obvious that would find a statistical significance anyways...
And yes, sample size doesn't matter because the significance already takes that into account. A test with a small sample will have a harder time getting statistically significant. However, if something is significant you have to recognize that the sample size was part of the calculation.
|
On September 03 2010 11:20 Tray wrote:Show nested quote +On September 03 2010 11:07 Acritter wrote:On September 03 2010 05:16 Acritter wrote:
t: 365200 36.6%
This is the percentage of Terran players in all of Starcraft 2. It is not even close to being accurately represented in upper Diamond League. Tray, please stop telling us all to compare this to data that has already been obtained and which contradicts your points. The fact remains that the high proportion of Terran players in upper Diamond cannot be dismissed as an outlier. The 'fact' is that it is dismissed as an outlier. Without any debate. Those numbers are not relevant. Use your value in the guy above me's calculation. You will see that you cannot say with a 95% certainty that they are overrepresented. Period. Goodbye idiots. I hope they lock this thread because people are only getting dumber coming here.
You have by far been the worst part of this thread.
Your condescending tone has done nothing but detract from your information.
Your posts are so thick with insults, ad hominem attacks and arrogance that it overshadows any information you tried to share.
Your time here was wasted by your own inability to discuss issues like an adult.
Learn to debate.
|
On September 03 2010 11:18 Tray wrote:
You're right, this is math. But this is not a correct analysis. You're actually arguing something completely different. I know you're not very smart so you don't probably know that. You probably think you compared the same thing as me.
Ad hominem is not necessary.
You're taking the expected number of terrans per player to be .3. Thus if there are 20 people you're saying that 6 of them are expected to be Terran. The actual number is 10 in this case so you're saying that since that's 4 points away from 6, it only has a 5% chance to be true.
Now lets take the real number of % of players who play Terran. 40%. Using your same analysis we would expect out of the top 20 people, 8 of them should be Terran. Now do your cumulative probability and tell me that it's still outside the norm.
Then there is a 76% chance that I am right.
But why should I use 40% rather than 30%? I used the percentage of non-terrible Diamond players. That is surely better than including all of the terrible bronze players who don't build more than 5 workers.
|
On September 03 2010 11:18 Tray wrote:Show nested quote +On September 03 2010 11:06 Sentient wrote:On September 03 2010 10:44 Tray wrote:
Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable.
As the sample drops, the error in the 'real' value increases and becomes less accurate. Period.
Stay in school. Tray, I'm going to show you are wrong beyond a reasonable doubt. I will use the calculator found here: http://stattrek.com/Tables/Binomial.aspxThis calculator does not rely on binomial distributions. It uses discrete calculations, just as the ladder behaves. From the spreadsheet linked earlier, there are 27862 players with 600 or more points in diamond. Terran makes up 30.6% of that. Thus we can treat the success of a Terran player as 30.6%, or .306%. In other words, if player skill is equally distribute (and you said yourself that we should assume this), 30.6% of the top 25 players should be Terran. So we plug .306 as the "Probability of success on a single trial", because this is the fraction of Terran players. Number of trials is 20, because that is the edge of the ladder ranking we are looking at. There are 10 Terran players in the top 20 as of right now (sc2ranks.com). This is the number of successes. Now press the calculate button. We are interested in Cumulative Probability: P(X >= 10). It will spit out 0.0544, or about 5%. In other words, there is a ~5% chance for 10 of the top 20 to be Terran. In other other words, there is a 95% probably that the distribution of Terrans in the top 20 is not due to chance. This is good enough for scientific publication. Sample size has nothing to do with it. Plug in a fake coin flip. Let's say our number events is 2, and we have 2 hits. Probability of success is 0.5. The odds of getting 2 (or more, but that's irrelevent here) is 0.25. Thus I could say with 75% certainty that the dice were rigged. I will be right 75% of the time and wrong 25% of the time. That's how it works. We're talking a lot more than 75% certainty. 95% certainty. The distribution of players falls outside what would be expected by chance, and sample size has nothing to do with it. (Ignoring the fact that this isn't technically a sample but is rather the complete reality.) There are only three reasonable explanations: 1. We are in the 5% zone, and this Terran distribution is really just a fluke. 2. Terran is overpowered. 3. Skilled players tend to play Terran more than the other races, and skill distribution is not equal (which was one of our key assumptions). This is math and it is fact. The Terran distribution falls outside random chance. I am 95% sure of this statement, which is a lot better than you can provide. You're right, this is math. But this is not a correct analysis. You're actually arguing something completely different. I know you're not very smart so you don't probably know that. You probably think you compared the same thing as me. You're taking the expected number of terrans per player to be .3. Thus if there are 20 people you're saying that 6 of them are expected to be Terran. The actual number is 10 in this case so you're saying that since that's 4 points away from 6, it only has a 5% chance to be true. Now lets take the real number of % of players who play Terran. 40%. Using your same analysis we would expect out of the top 20 people, 8 of them should be Terran. Now do your cumulative probability and tell me that it's still outside the norm. There I just used your own analysis to prove you wrong. Feel free to do me now with mine from before.
Hey genius, since you clearly have so much time on your hands that you can reply to nearly every single post with some random nitpicking, and you are obviously so well versed in statistics, do the math then, for 1200+, 1300+, 1400+ or 1500+ with your "40%" number (source?) and post it. Otherwise, all you're doing is tossing out ad-hominems and generally being a dick.
|
On September 03 2010 10:44 Tray wrote:Show nested quote +On September 03 2010 10:40 blacktoss wrote:Yes I can claim that. It's EXACTLY what I'm claiming and it's true. The person you're referring to did the worst math and completely ignored the fact that Terran isn't expected to be 33% representative. It should be relative to the total population of players that play terran if skill is evenly distributed between the three races, which yes, we can assume.
So his one in a billion was not even remotely accurate. In fact these numbers near the tail end are without any doubt are statistically insignificant. He didn't lie, he just didn't know what he was doing. Here's the kicker genius: You're fighting a straw man. The claim made was that Terran make up a disproportionate number of high level diamond players such that the deviation from an even distribution is not due to random chance. The claim is CORRECT. The math is CORRECT. You are the one coming in here and putting words in peoples' mouths and then saying "haha no u r rong I will now insult you". I don't think you know how statistics works at all. The hypothesis tested was the hypothesis that Terran should make up 33% of the racial distribution in high diamond league (I am not sure if he took into account Random). The statistical test he used showed with high confidence that this hypothesis was false. End of story that is all the reasoning used. You talk about statistical insignificance but your criticism does not address the point's validity, it attacks it on the basis that "it is not the right hypothesis". Sorry, but that does not say anything about the conclusion drawn. Maybe you should just look it up for yourself and get smarter instead of being like everyone else in here and demanding I teach you all how statistics work.
This is 5 seconds of google.
A typical example would be when a statistician wishes to estimate the arithmetic mean of a quantitative random variable (for example, the height of a person). Assuming that they have a random sample with independent observations, and also that the variability of the population (as measured by the standard deviation σ) is known, then the standard error of the sample mean is given by the formula:
σ/sq.root(n)
As N becomes smaller and smaller, where N is the population of the sample size, the error increases.
if n = 25, error = std/5 (std = 1) = 20%
if n = 50, error = std/7.07 = 14%
if n = 100, error = std/10 = 10%
if n = 1000, error = std/31.6 = 3% Once again, you are full of shit. You come in here saying you are better than anyone else, throw around a few big words, and then cite google and say "hurdy hur". It doesn't matter if the variance is high, because when YOU DO THE ACTUAL TEST (instead of bullshitting with jargon), you find that p is very high. So high in fact that you MUST discard the hypothesis that "Terran is not overrepresented in high diamond". You can say "WELL SIGMA IS HIGH" but it doesn't matter. Statistical tests are robust. They take sigma into account. I am beginning to feel like you don't know anything about statistics at all. Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable. As the sample drops, the error in the 'real' value increases and becomes less accurate. Period. Stay in school.
PRECISELY. you have to take into account the number of players for each race.
plus, aside all this BS about whether its imba or not and how we can or cannot show it mathematically. has anyone though blizzard's system might be IMBA or broken. its a VERY aribitrary way of assigning "skill", points, or whatever. might one ask a totally different question, that is, "is blizzards ranking system even set up such that in ideal conditions it will yield a gaussian distribution?"
|
I would like to further add this note to Tray.
On September 03 2010 11:01 Tray wrote:
These numbers take into account the total race distribution relative to the % of people that play that race. This however does not include the error.
To find the error simply take the equation I just gave, 1/ sqroot(n) to find the error in those numbers. Now I forget from here if you would have to divide it by 3, or do a more complicated statistical analysis to figure out how much error applies to each normalized race %, but I'm sure if you read the wikipedia page on sampling you can figure out the right way to account for it.
It doesn't matter though because it's obvious from how much the error increases as the sample drops that the 'real' numbers will vary so much at the low population it'll look something like this (with 95% confidence interval); T = 10-50%, Z = 5-35%, P = 10-50%. Very obviously not reliable data because of the huge volitility. More complicated analysis can be done, but it's not necessary.
This is a faulty analysis of confidence because the confidence is already 100%. This is a dataset where n = 1. The top 20 players are the top 20 players, and the top 100 players are the top 100 players.
Your analysis would be correct if we were talking about a random 20 or 100 sampling from the entire pool of players, but we aren't. We are discussing a well-defined set of players, and there is no error or uncertainty involved.
This is a common mistake people keep making. It is inappropriate to talk about sample size when discussing the top players, because the set of top X players is not a random sampling of a larger population. Again, there is no error or uncertainty in the top 20 or top 100 or top 5000. These are discrete numbers and they are what they are.
|
On September 03 2010 11:54 yourwhiteshadow wrote:Show nested quote +On September 03 2010 10:44 Tray wrote:On September 03 2010 10:40 blacktoss wrote:Yes I can claim that. It's EXACTLY what I'm claiming and it's true. The person you're referring to did the worst math and completely ignored the fact that Terran isn't expected to be 33% representative. It should be relative to the total population of players that play terran if skill is evenly distributed between the three races, which yes, we can assume.
So his one in a billion was not even remotely accurate. In fact these numbers near the tail end are without any doubt are statistically insignificant. He didn't lie, he just didn't know what he was doing. Here's the kicker genius: You're fighting a straw man. The claim made was that Terran make up a disproportionate number of high level diamond players such that the deviation from an even distribution is not due to random chance. The claim is CORRECT. The math is CORRECT. You are the one coming in here and putting words in peoples' mouths and then saying "haha no u r rong I will now insult you". I don't think you know how statistics works at all. The hypothesis tested was the hypothesis that Terran should make up 33% of the racial distribution in high diamond league (I am not sure if he took into account Random). The statistical test he used showed with high confidence that this hypothesis was false. End of story that is all the reasoning used. You talk about statistical insignificance but your criticism does not address the point's validity, it attacks it on the basis that "it is not the right hypothesis". Sorry, but that does not say anything about the conclusion drawn. Maybe you should just look it up for yourself and get smarter instead of being like everyone else in here and demanding I teach you all how statistics work.
This is 5 seconds of google.
A typical example would be when a statistician wishes to estimate the arithmetic mean of a quantitative random variable (for example, the height of a person). Assuming that they have a random sample with independent observations, and also that the variability of the population (as measured by the standard deviation σ) is known, then the standard error of the sample mean is given by the formula:
σ/sq.root(n)
As N becomes smaller and smaller, where N is the population of the sample size, the error increases.
if n = 25, error = std/5 (std = 1) = 20%
if n = 50, error = std/7.07 = 14%
if n = 100, error = std/10 = 10%
if n = 1000, error = std/31.6 = 3% Once again, you are full of shit. You come in here saying you are better than anyone else, throw around a few big words, and then cite google and say "hurdy hur". It doesn't matter if the variance is high, because when YOU DO THE ACTUAL TEST (instead of bullshitting with jargon), you find that p is very high. So high in fact that you MUST discard the hypothesis that "Terran is not overrepresented in high diamond". You can say "WELL SIGMA IS HIGH" but it doesn't matter. Statistical tests are robust. They take sigma into account. I am beginning to feel like you don't know anything about statistics at all. Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable. As the sample drops, the error in the 'real' value increases and becomes less accurate. Period. Stay in school. PRECISELY. you have to take into account the number of players for each race. plus, aside all this BS about whether its imba or not and how we can or cannot show it mathematically. has anyone though blizzard's system might be IMBA or broken. its a VERY aribitrary way of assigning "skill", points, or whatever. might one ask a totally different question, that is, "is blizzards ranking system even set up such that in ideal conditions it will yield a gaussian distribution?"
Your question doesn't make sense. Please revise it?
|
He wants the distribution of player points to follow a bell curve. That is, about 2/3 of players fall within one standard deviation of points, 95% of players fall within 2 standard deviations, etc.
The correct answer is: No, I would hope that Blizzard's system is not set to yield a Gaussian distribution. Player skill is not Gaussian, as shown by extensive analysis of chess games.
|
sorry, "is blizzards ranking system even set up such that in ideal conditions (a perfectly balanced game where only skill is a determining factor in wins/losses) it will yield a gaussian distribution?"
we're talking about a game that is supposed to bring in revenue for blizzard. if everyone had the perception that they are average, would they continue to play the game? let's say the avg at so and so month is 600 pts, do you want to be an average diamond playaer? HECK NO, you're in diamond. if you were posting about strategy, you would start off with: "hi i'm a xxx pts INSERT RACE player". what i'm trying to say is, that blizzard might intentionally want to pwn the distribution, or it might just be inherently retarded because blizzard sucks at making a ladder. also, like someone else said, this ladder thing isn't really indicative of skill. you can mass games and be at the top, i mean there is even a BONUS pool.
|
Aotearoa39261 Posts
Troll eliminated. Discussion, please continue.
|
aww troll is gone  while i wanted to side with his arguments, his trolling was just...trollerific
|
Thank you, Plexa. Your efforts are much appreciated.
So, uh, does anyone remember where we were before Tro- I mean Tray showed up?
|
the last 2 sections of the graph are pretty lol :D
interesting how many protoss in the mid-level ranges
|
So sad to see all the people arguing "statistical significance" with that horrible troll for 10 pages
alone the data is questionable as to its significance yes
but its not alone is it.....?
its paired with hundreds of pages of anecdotal evidence,
hundreds of hours of personal experience playing and watching games
and multiple credible expert sourse indicating there is some imbalance......
so to argue that it carries no significance, statistical or otherwise is incredibly Naive or deliberately misleading
also i didn't like the trolls attitude...so arrogant and full of personal attacks, that is completely unnecessary
good job mods
|
So, I've actually been working on a similar analysis for the past few days. I'm going to bring up a few points:
1) It's entirely possible for the "sample size" to be too small, even if it represents the entire population. If you are trying to find a correlation between two variables (race preference and rank, in this case), you have ventured into the mystic realm of inferential statistics. Many inferential methods yield poor results if there aren't enough data points. (This was a major problem in my initial analysis attempts, because the low number of random players was causing problems with the method I employed).
2) I disagree with your method of grouping players by point range. It makes the upper ranges less reliable (due to smaller population, as described above), and that's the most relevant players. I chose to separate players into bins of 50, sorted by rank (as pulled from SC2Ranks). That netted me a graph like this:
![[image loading]](http://img245.imageshack.us/img245/4017/racialpreferencebybin.png)
As you can see, there is still a possible favoring of Terran, but it's not quite so smooth. Is it just a coincidence that all those Terrans are in the top bin?
3) There are actually ways to mathematically calculate the likelyhood of this. I chose to use something called the chi-square goodness of fit test, which essentially tests whether a population matches an expected ratio. Testing each bin, I created the graph below:
![[image loading]](http://img405.imageshack.us/img405/6385/pvalues.png)
The red line marks the threshold of statistical significance. As you can see the top and bottom bins do deviate from the average ratio of the whole in a significant way. This is exactly what the graph should look like if some races perform better than others.
4) So this proves Terran are imbalanced? Unfortunately, no. The method I used has some problems with it.
First, it doesn't specify which races are imbalanced.
Second, it is possible that Random players are "imbalanced," in the sense that players who focus on a specific race play better than players who focus on all races. In that case, such an "imbalance" would not reflect an actual problem with gameplay. See that significant result in bin 12? That's because there is a group of four random players in it. It doesn't take many to skew the results.
Third, this reflects only one region (North America) at only one point in time (two days ago, when I pulled this data from SC2Ranks). If further analysis showed that margin growing or shrinking, or showed different results in different regions, then a fundamental imbalance in gameplay would be unlikely.
I am working on this at least a few hours a day, and I hope to have something more conclusive posted this weekend.
-The Scarlet Mathematician
|
On September 03 2010 10:44 Tray wrote:
Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable.
I know with 95% certainty that in an internet argument, the guy who is not throwing around personal insults is right and the guy who is doing that is wrong.
By the way, please consider what the reasons might be that so many highly skilled players want to use Terran when they are playing to win, and don't choose to use Zerg. (And if you think it's because of the campaign, you must be joking.)
On September 03 2010 10:56 Tray wrote:
This is not a correct statement. Blizzard is not balancing the game around statistically insignificant data.
Likely they are using data of a majority of Diamond and then looking at individual games from tournaments and very top ranked players to try to manually spot what we refer to as 'cheese' and 'abuse.' This is because the game is evolving and probably does so from the top down.
LOL, they're nerfing BCs and ultras and leaving marauders alone. Blizzard's balancing techniques involve dartboards and monkeys without a doubt.
|
U guys are reading this completely wrong...
Zerg is underpowered and needs fixed. Thats pretty well the only thing that is conclusive from this chart.
|
Valhalla18444 Posts
On September 03 2010 16:00 Scarmath wrote:
post
oh wow... please update ASAP
|
Maybe this has been provided already, but where can I get the player data in a csv file? I'd like to play around with some tests as well
|
On September 03 2010 11:59 blacktoss wrote:Show nested quote +On September 03 2010 11:54 yourwhiteshadow wrote:On September 03 2010 10:44 Tray wrote:On September 03 2010 10:40 blacktoss wrote:Yes I can claim that. It's EXACTLY what I'm claiming and it's true. The person you're referring to did the worst math and completely ignored the fact that Terran isn't expected to be 33% representative. It should be relative to the total population of players that play terran if skill is evenly distributed between the three races, which yes, we can assume.
So his one in a billion was not even remotely accurate. In fact these numbers near the tail end are without any doubt are statistically insignificant. He didn't lie, he just didn't know what he was doing. Here's the kicker genius: You're fighting a straw man. The claim made was that Terran make up a disproportionate number of high level diamond players such that the deviation from an even distribution is not due to random chance. The claim is CORRECT. The math is CORRECT. You are the one coming in here and putting words in peoples' mouths and then saying "haha no u r rong I will now insult you". I don't think you know how statistics works at all. The hypothesis tested was the hypothesis that Terran should make up 33% of the racial distribution in high diamond league (I am not sure if he took into account Random). The statistical test he used showed with high confidence that this hypothesis was false. End of story that is all the reasoning used. You talk about statistical insignificance but your criticism does not address the point's validity, it attacks it on the basis that "it is not the right hypothesis". Sorry, but that does not say anything about the conclusion drawn. Maybe you should just look it up for yourself and get smarter instead of being like everyone else in here and demanding I teach you all how statistics work.
This is 5 seconds of google.
A typical example would be when a statistician wishes to estimate the arithmetic mean of a quantitative random variable (for example, the height of a person). Assuming that they have a random sample with independent observations, and also that the variability of the population (as measured by the standard deviation σ) is known, then the standard error of the sample mean is given by the formula:
σ/sq.root(n)
As N becomes smaller and smaller, where N is the population of the sample size, the error increases.
if n = 25, error = std/5 (std = 1) = 20%
if n = 50, error = std/7.07 = 14%
if n = 100, error = std/10 = 10%
if n = 1000, error = std/31.6 = 3% Once again, you are full of shit. You come in here saying you are better than anyone else, throw around a few big words, and then cite google and say "hurdy hur". It doesn't matter if the variance is high, because when YOU DO THE ACTUAL TEST (instead of bullshitting with jargon), you find that p is very high. So high in fact that you MUST discard the hypothesis that "Terran is not overrepresented in high diamond". You can say "WELL SIGMA IS HIGH" but it doesn't matter. Statistical tests are robust. They take sigma into account. I am beginning to feel like you don't know anything about statistics at all. Dude you are incredibly stupid. My analysis is correct. Assuming Terran should make up 33% of the distrubtion at the top is NOT correct. It should be EQUAL to the total % of players that play terran. Period. It's not debatable. As the sample drops, the error in the 'real' value increases and becomes less accurate. Period. Stay in school. PRECISELY. you have to take into account the number of players for each race. plus, aside all this BS about whether its imba or not and how we can or cannot show it mathematically. has anyone though blizzard's system might be IMBA or broken. its a VERY aribitrary way of assigning "skill", points, or whatever. might one ask a totally different question, that is, "is blizzards ranking system even set up such that in ideal conditions it will yield a gaussian distribution?" Your question doesn't make sense. Please revise it?
Don't bother that argument is pointless anyways. Because there is no sample size. It's the entire population above a certain pt level. Therefor sigma is zero. There is no margin of error becuase you have taken the whole population, there is no more ppl you can add to the sample to change the results.
In case anyone else doesn't know sigma represents the % your results could be diffrent from the actual results if you had the entire population. If the sample = population then sigma is 0.
|
On September 03 2010 20:55 Flyingdutchman wrote:
Maybe this has been provided already, but where can I get the player data in a csv file? I'd like to play around with some tests as well
first page. ctrl+f "csv" you'll see it.
|
@Scarmath I like what you have done, still I got some questions.
What exactly do the bins represent? Have you crawled only the Diamond ladder players, and where is your cut if you have done so - or did you read out the ranking from position 1 to 1000 grouping the results per 50 each into a bin giving you 20 bins?
Did you include the high point low league players, too?
I am also not sure how to interpret the graph, since it still is discret and you show us curves, have you run a density approximation over it or is it polynomial fitting? Since it ends right in the middle of the first bin I think the graph itself represents a smoothed curve fitting.
What exactly have you done with the Chi Square test (or better if you want you can post your data as csv - or give a timestamp)?
What do you think is needed to solve the random problem? We could do a very bad approximation and tell that random is 33% Terran, Zerg, Protoss (leading to some minor errors, aka it would screw some findings because gain and loss of a race would not be taken into account equally).
I've read that you only watched the NA server, this may further alter your results in comparision with mine. Even since Terran is overrepresented in Korea there are some more Zerg than in the US.
Still I think the ultimate findings and your tests give some nice results (even if I think the random problem has to be solved), about general non diamond league distribution.
I'm looking forward to see your later findings.
*Edit format and made a sentence readable*
|
On September 03 2010 22:01 yourwhiteshadow wrote:Show nested quote +On September 03 2010 20:55 Flyingdutchman wrote:
Maybe this has been provided already, but where can I get the player data in a csv file? I'd like to play around with some tests as well first page. ctrl+f "csv" you'll see it.
thanks, but the data isn't exactly in the format I was looking for. I was looking for an export of the entire diamond ladder, either over all regions, but grouped into different regions would be better. The csv in the OP just has the total number of players of each race in the different groupings (900-1000; 1001-1100; etc). Does anybody know of a way to pull the ladderdata into a csv? like player A, x1 points, player B, x2 points?
edit: and the data should of course include the preferred race of the player
|
On September 03 2010 16:00 Scarmath wrote:3) There are actually ways to mathematically calculate the likelyhood of this. I chose to use something called the chi-square goodness of fit test, which essentially tests whether a population matches an expected ratio. Testing each bin, I created the graph below: The red line marks the threshold of statistical significance. As you can see the top and bottom bins do deviate from the average ratio of the whole in a significant way. This is exactly what the graph should look like if some races perform better than others.
Can you please give more detail how you did this? Depending on how you treat the data, a chi-square test may or may not be appropriate. I'm also interested in how you combined the fits for the different races.
I think it is incorrect to say the OP's binning method "makes the upper ranges less reliable (due to smaller population, as described above)". There is no reliability to the rankings -- they are what they are. It is not a random sampling of the population, it is the population.
Perhaps you could do your same analysis by running a sliding window of 20 players over the entire population? I think this would yield a very interesting graph. If you do this, you would expect about 1 in 20 points to fall beyond statistical significance; if more than 1 in 20 do at the high range, then there is truly something fishy going on. I'd also be happy to do it myself, but it looks like you already have the spreadsheet to do it easily.
|
On September 03 2010 22:28 ReplayArk wrote:
@Scarmath I like what you have done, still I got some questions.
What exactly do the bins represent?
The bins are segments of 50 players each. Bin 1 is ranks 1-50, bin 2 is ranks 51-100, etc.
Have you crawled only the Diamond ladder players, and where is your cut if you have done so - or did you read out the ranking from position 1 to 1000 grouping the results per 50 each into a bin giving you 20 bins?
I basically extract the data from the ranking tables on SC2Ranks.com. There is nothing stopping me from doing more or less, but 1000 seemed like a nice starting point.
Did you include the high point low league players, too?
No, that was outside the scope of what I was trying to do.
I am also not sure how to interpret the graph, since it still is discret and you show us curves, have you run a density approximation over it or is it polynomial fitting? Since it ends right in the middle of the first bin I think the graph itself represents a smoothed curve fitting.
I smoothed the lines to make it pretty. (I've actually stopped doing this on my more recent graphs, but this was the one I had finished).
What exactly have you done with the Chi Square test (or better if you want you can post your data as csv - or give a timestamp)?
The Chi-Square test I used basically tests if an observed population distribution matches a hypothesized distribution. In this case I hypothesized that each bin's distribution matched the overall distribution.
I'll see if I can post the spreadsheet somewhere.
What do you think is needed to solve the random problem? We could do a very bad approximation and tell that random is 33% Terran, Zerg, Protoss (leading to some minor errors, aka it would screw some findings because gain and loss of a race would not be taken into account equally).
Either by testing each race preference individually (i.e. testing the success of Terran vs non-Terran), or by finding a test that can leave out random. I have three or four things I'm looking at, I just need time to slog through the math to make sure they're applicable.
I've read that you only watched the NA server, this may further alter your results in comparision with mine. Even since Terran is overrepresented in Korea there are some more Zerg than in the US.
Still I think the ultimate findings and your tests give some nice results (even if I think the random problem has to be solved), about general non diamond league distribution.
I'm looking forward to see your later findings.
Yes. I think, ultimately, Diamond is the best place to analyze balance because there is no way for good players to be promoted out of it. A comparison between Platnium and Diamond might be interesting, though.
|
Can you please give more detail how you did this? Depending on how you treat the data, a chi-square test may or may not be appropriate. I'm also interested in how you combined the fits for the different races.
I think I answered this above.
I think it is incorrect to say the OP's binning method "makes the upper ranges less reliable (due to smaller population, as described above)". There is no reliability to the rankings -- they are what they are. It is not a random sampling of the population, it is the population.
To take, as an example, the chi-square test, it relies on never expecting a population of zero, and usually expecting a population more than 5. If I set my bin size to 10 instead of 50, in many cases the expected number of random players is less than one. The rankings "are what they are," but the number of people affects how confidently we can say anything about them. Binning by point range means that you have very small bin sizes at the top, which will sap the reliability of analyzing them. That's why I prefer static bin sizes.
Perhaps you could do your same analysis by running a sliding window of 20 players over the entire population? I think this would yield a very interesting graph. If you do this, you would expect about 1 in 20 points to fall beyond statistical significance; if more than 1 in 20 do at the high range, then there is truly something fishy going on. I'd also be happy to do it myself, but it looks like you already have the spreadsheet to do it easily.
A bin size of 20 is too small (even a bin size of 25 is pushing it), but I did try this with a sliding window of 50. It was...weird. Actually, hold on, let me whip up a graph of it, so you can see what I mean...
EDIT:
...OK, that didn't take as long as I though. Here is the racial preference by rolling bin chart:
![[image loading]](http://img696.imageshack.us/img696/1910/racialpreferencebyslidi.png)
and here are the corresponding p-values:
![[image loading]](http://img15.imageshack.us/img15/6199/angrypvalues.png)
The graph actually says pretty much the same thing. I went back to static, separate bins, though, because I felt it was easier to visually parse.
|
haha oh my god, thats pretty awesome 6:1 T : Z
|
On September 04 2010 05:47 Scarmath wrote:
To take, as an example, the chi-square test, it relies on never expecting a population of zero, and usually expecting a population more than 5. If I set my bin size to 10 instead of 50, in many cases the expected number of random players is less than one. The rankings "are what they are," but the number of people affects how confidently we can say anything about them.
This is where I have to disagree. The chi-square test tells you exactly how confident you can be, regardless of the bin size. That's the whole point of using the p-values. The smaller bin sizes are accounted for by the t-table, so it doesn't matter how big or small they are. Smaller bins will use a larger uncertainty multiplier.
A bin size of 20 is too small (even a bin size of 25 is pushing it), but I did try this with a sliding window of 50.
What do you base this assertion on? There are tests to establish these probabilities. Everyone keeps saying that 20 is too few, yet no one has actually done that calculation.
and here are the corresponding p-values: The graph actually says pretty much the same thing. I went back to static, separate bins, though, because I felt it was easier to visually parse.
This says a lot actually. With only random chance, we expect 1 in 20 of the data points to be p < 0.05. That's about what you have. However, there is a large cluster at the top 100 players where p < 0.05, and this stays true regardless of which 20 of the top 100 you pick. This is exactly what you would expect if racial distribution was not due to chance. Therefore, we can say with very good confidence (~95%) that the racial distribution is not caused by random chance. It doesn't tell us anything about the reasons for that, but it does give us a hard confidence number of that fact.
If you're still concerned about the bins, could you post your spreadsheet somewhere? I'd like to play around with it. It seems like you have it set up in a very convenient way to manipulate.
|
This is where I have to disagree. The chi-square test tells you exactly how confident you can be, regardless of the bin size. That's the whole point of using the p-values. The smaller bin sizes are accounted for by the t-table, so it doesn't matter how big or small they are. Smaller bins will use a larger uncertainty multiplier.
and
What do you base this assertion on? There are tests to establish these probabilities. Everyone keeps saying that 20 is too few, yet no one has actually done that calculation.
Yes, I've done calculations with bin sizes of 10 and 25, and neither provided clear results (at least to my satisfaction). The problem is that with so few random players, the presence or absence of merely one or two random players will create a false positive. Here, I was trying to avoid this, but let's look at the formula I was using. (Math-o-phobes should shield their eyes).
![[image loading]](http://img716.imageshack.us/img716/1937/formula.png)
where i iterates over the the categorical groups G (race preferences, in this case), O is the observed count, and E is the expected count.
The first thing to note is that if E is zero, we get a divide by zero error. Furthermore, for small values of E, relatively small perturbations in O can result is overly large chi-square scores, which in turn can lead to false positives. Why is this a problem? Because random players make up around 5% of the top 1000. With a bin size of 20, that gives us an expected value of 1 per bin, which, in turn, gives the presence of random players a larger influence on the final p-value. A bin size of 50 gives an expected value of 2.85, which is still very low, but at least lessens the impact of Random players. (My statistics text suggests the majority of your expected values be above 5, but I have no idea where they got this number).
Also...
This says a lot actually. With only random chance, we expect 1 in 20 of the data points to be p < 0.05. That's about what you have. However, there is a large cluster at the top 100 players where p < 0.05, and this stays true regardless of which 20 of the top 100 you pick. This is exactly what you would expect if racial distribution was not due to chance. Therefore, we can say with very good confidence (~95%) that the racial distribution is not caused by random chance. It doesn't tell us anything about the reasons for that, but it does give us a hard confidence number of that fact.
Yeah, but that's also exactly what my other graph said. (That is, the top 100 players where statistically significantly different than the overall distribution of the top 1000). The issue is determining if that is because of random players being underrepresented, or zerg players being underrepresented.
If you want my (unsubstantiated) opinion, I think that it is because zerg players are underrepresented. However, without more refined methods, we just can't jump to that conclusion. (At the very least, we can't use numbers as proof).
|
Ok, I understand now. Thank you. I hadn't considered the effect of random players (because who cares about random players anyway ?).
When I came up with my 95% figure earlier, I tested the more specific hypothesis that Terran is overrepresented. This is simpler and more to the point than looking at the distribution of all four options. Since Terran make up 30-40% of the population (depending on which subset you take), 20 is more than large enough for a bin size.
|
United Arab Emirates492 Posts
Sentient and Scarmath can you guys post your findings in laymen terms . It is really difficult for me (and I am sure most of the viewers) to decode and understand stuff from your charts.
Edit: Post -> Most obviously.
|
On September 04 2010 11:27 Gunman_csz wrote:Sentient and Scarmath can you guys post your findings in laymen terms . It is really difficult for me (and I am sure post of the viewers) to decode and understand stuff from your charts.
I skimmed mostly but the jist of it is theyre trying to account for the affect random players have on the distributions. Theres also some attempts to adjust for a fact that there is not an equal distribution of population in all racial choices (t/p/z/r).
Theres a lot of great math in here and as an actuary I'm really surprised and impressed this is out on a video game board.
Unfortunately theres too much missing information to make any of this useful. Taking a step back from the math look at the system youre working. More specifically the match making system. How does it work? From the visible system youre matched with an opponent and you win or lose points based on their HIDDEN elo point system. Does a player ever get matched with a different league/large visible point discrepancy, yes all the time infact theres always whine on the blizzard forums about this.
Are points only awarded for skill? Negative only wins and losses. For example can we agree that idra is one of the best sc2 players right now? I mean hes favored to win the GSL, he also has a paltry ~1300 score (rounded up) Why is that? Idra doesnt ladder a lot infact the whole system as seen by us rewards those who play a lot. How many times have you played a terrible opponent only to look him back up and seen an incredible amount of points with a terrible record? Further more how can idra clearly not be in the top 100 based on visible points but still be on blizzards top 100? Blizzard internally uses a completely different system to check players determined ELO for balance purposes. One not skewed by point inflation for mass gaming.
I could go on but this is the main point is this, the graph and subsequent math adventures is just a graph of the number of people in xyz visible points that are in no way correlated to any sort of balancing information, or atleast not any useful information. Its very much of a representation of who has the most cookies in an imaginary jar (assuming there is actually a jar of cookies divided on a completely different system than what is shown)
Finally: This is not the metric for blizzards balancing FYI nor should it ever be the game would be terrible.
|
On September 04 2010 11:27 Gunman_csz wrote:Sentient and Scarmath can you guys post your findings in laymen terms . It is really difficult for me (and I am sure post of the viewers) to decode and understand stuff from your charts.
TL;DR version: The racial distribution in the top 100 players is not expected by random chance. The two probable reasons are (1) The races are not equally powerful, and/or (2) Player skill is not equally distributed among the races.
Scarmath did all of the graph work. I mostly plugged some numbers into a binomial probability calculator (http://stattrek.com/Tables/Binomial.aspx). He did a lot more work and should get the credit.
Imagine you want to know if a dice is weighted. You roll it 10 times and get 5 sixes. Knowing that the probability of rolling a six is 1 in 6, the odds of rolling a six 5 or more times out of 10 is 0.07%. With this information, we can conclude with 99.93% certainty that the dice is weighted.
The point of contention is that 10 isn't a very large dataset. Say we are interested in how many times we roll a 1, which happens to be zero. There is a 24% chance of this occurring by chance. One could argue with 76% certainty that this is not due to chance.
Since there are six possible outcomes in a dice roll, chances are that at least one of the six numbers will come up 0 times. This means that you have to be very specific in the question you are asking about the dice, since it's easy to cherry pick the statistics in your favor. I could choose any of the six numbers with an above- or below-average abundance to argue my point, but this would be dishonest, because I am looking for data to support my preconceived notion that the dice is weighted.
We can use the same statistical means to test the racial distribution in the top 100 players. Knowing the distribution of players in Diamond league (eg, 30.6% Terran, etc.), we can calculate the probability of the top x players' races occurring purely by random chance. For example, we would expect 30.6% of the top 100 players to be Terran.
In reality, Terran are more abundant than 30.6% in the top 100.
-Is this caused by random chance?
-Is it caused by only looking at 100 players?
The answers to both of these are no. Using either of our analyses (mine using a binomial calculator, Scarmath's using a chi-square test), we both find that there is a 95% chance that the distribution of players in the top 100 cannot be explained by random chance. This is an all encompassing number, independent of the number of players involved. Whether we looked at 5 or 20 or 1000 players is irrelevant, because that 95% figure incorporates the number of players.
As with all analyses, this has to be combined with what we already know about the game. All it tells us is that random chance has not produced a high ratio of Terrans in the top 100. Other factors must come into play. For example,
- Random players are less abundant in the top of the ladder, likely because it becomes harder and harder to compete as random. Random players must learn 9 matchups instead of 3.
- The "Terran's OP" meme probably shifts the player distribution in favor of Terran.
The 95% figure doesn't tell us that Terran is OP, that Zerg is UP, or that Mars is really blue. All it tells us is that something other than chance (or "sample sizes" for that matter) must be used to explain the proportion of Terran players in the top 100. This is true regardless of your stance on the Terran issue. What to do with this information is another story, and it will only shift the flamewar back to whether or not Terran is balanced.
Lies, damned lies, and statistics indeed.
|
I read the first 6 pages and all the pseudo-statistics almost made me cry.
I have a look at page 19 and 20 and apparantly no improvement. So it's time to enter this topic i'm afraid:
As several people noted, the data contains the whole population of diamond 600+ players.
Since you know the entire population, there is no need at all for statistics of significance.
Statistics are what you use when you only know part of the population. In that case, there is uncertainty and statistics helps to make statements about the whole population, based on the observed sample, with a specific level of certainty.
Every difference you see, is a real, existing difference. Any equality you see, is a real, existing equality.
The question is not whether terran is more played on higher skill levels, we know that since we know that in the total population of the top scoring players, more than 1/3 plays terran.
The questions left to answer are:
1. Are the differences so big that we need to act?
2. What are the causes of these differences?
Any test of significance in this topic is used inappropriately and worthless. Unless you would add the finite population correction. But that would only prove my earlier point that any difference is a real difference, since the FPC part in this case would always be 0, meaning your confidence interval will always have a width of 0, so anything that has even the slightliest difference, would be significant.
That being said, please stop posting all the worthless pseudo-statistical crap.
|
@Shorack you may just read the first post and the edits, it will lead you to some discussion from Scarmath and Sentient and what they have doone is not worthless pseudo-statistical crap. You may read it if you want to discuss anything further, but it is not mannered to implicitly tell the other they are dumb, or else why should one write worthless pseudo-statistical crap?
|
On September 04 2010 23:55 ReplayArk wrote:
@Shorack you may just read the first post and the edits, it will lead you to some discussion from Scarmath and Sentient and what they have doone is not worthless pseudo-statistical crap. You may read it if you want to discuss anything further, but it is not mannered to implicitly tell the other they are dumb, or else why should one write worthless pseudo-statistical crap?
First of all, apologies for the use of the word crap. I got quite nervous from seeing people throw around statistics where they aren't appropriate. It's even painful to see that some people put so much effort in it, since it's inappropriate use and hence wasted effort.
Second, i went through the edits:
blacktoss is right, there is no confidence to consider.
too bad that he ignores that himself in the same post. since he refers to a chi-square test, which implies he wants to see if something falls within a confidence interval or not.
cotonou is a 100% right. note that he refrains from statistical tests (well, except for the use of the term null hypothesis )
toxigen's post is again an interesting one. note again that he doesn't try to use statistical stuff. (i hope you start to see the pattern in what i like and what not? )
on scarmath then: he did put quite some effort in it, but it's not because you put effort in something, that it's correct. Since we're observing the whole population, the differences are always significant. So using a chi-square test has no point, it's always significant unless the expected and observed are exact. (it's not the case in scarmaths second graph because he didn't correct for finite population)
I'll try to repeat myself in an attempt to make it clearer compared to my first attempt: there is no question of significance, every difference is significant. There is only a question of relevance (are the differences we see big enough to act upon?) and for the relevant ones: what's causing it.
Significance is about: is the difference we see, really there? And we have to ask that ourselves when only seeing part of the population. If you see the whole population, you see the differences that are really there and so no longer need to wonder about that, you know it for certain.
We still need to ask ourselves the relevance question: the difference we see, is it big enough to worry about (say we find there are 2% more terran players than expected, is it really worth the effort to work that away?) and finally, the why-question: what is the cause of the difference? (no point in changing game balance if that's not causing the difference)
I hope this makes my statement a bit clearer to you? (i fear my english is falling a bit short right now)
|
I'll try to repeat myself in an attempt to make it clearer compared to my first attempt: there is no question of significance, every difference is significant. There is only a question of relevance (are the differences we see big enough to act upon?) and for the relevant ones: what's causing it.
I think, perhaps, you don't understand exactly. Let me state all this explicitly.
It is clear that the proportions of the top 100 players are not equal to the proportions of the top 1000. The question is, how likely is that to be the result of random variation? We're not measuring sampling error. We don't need to measure sampling error. We do need to measure confidence.
That is the point of both the binomial calculation and the chi-square test. We already know there is a difference, but we need to test how statistically significant that difference is. In both cases, we can say with 95% certainty that the proportions of the top 100 players are not different due to random chance. That is, the deviation is statistically significant.
These are steps we need to take even when we have the entire population.
|
I've uploaded a version of the spreadsheet I've been working on here:
http://www.mediafire.com/?uta56tbvr5w8ect
The easiest way to play with it is by replacing the Raw table with other information copy and pasted from SC2Ranks.com. (Copy and paste it from Firefox. Chrome doesn't copy the Alt-test I use to identify race, and I haven't tested it in Internet Explorer). The rest of the sheets should update automatically.
Other things may be messed with, but may require more extensive fiddling to work.
Still working on this a few hours at a time.
|
On September 04 2010 23:24 Shorack wrote:
Since you know the entire population, there is no need at all for statistics of significance.
.
Absolutely not true. What could be happening in the population could be random chance. If you created a population randomly with balanced races, you won't get 33% 33% 33% (forget random to make it easier), you get something slightly off. Statistical significance tells you "if the races were balanced, there would be a P percent chance of this happening". When you get a number like .01%, you go "wow, there is almost no chance this randomly occurred".
However, if you look at the data, you can make an inference from something that is random. For example, if you flipped a coin 3 times, and you saw all heads, if you looked at the graph you'd go "wow, this is definitely heads biased." Statistics would tell you though "hey dude, chill. There is a 12.5% chance of that happening randomly, so I wouldn't jump to conclusions just yet".
That is why we are using statistics.
|
I would like to see some data compiled about ladder over time so maybe we can do time series analysis. Of course, time series are a bit more complicated than just a single sample/population.
|
On September 05 2010 04:52 rackdude wrote:Show nested quote +On September 04 2010 23:24 Shorack wrote:
Since you know the entire population, there is no need at all for statistics of significance.
. Absolutely not true. What could be happening in the population could be random chance. If you created a population randomly with balanced races, you won't get 33% 33% 33% (forget random to make it easier), you get something slightly off.
In the long run you would. After thousands of games, would you call the differences in that graph slight?
Use your noggin.
|
wont these stats be messed up no matter what due too the imbalance?
if you have two people and lets say their skill is 1000 and one of them plays terran and the other plays zerg, its fair too say the terran player would end up having a higher rating than the zerg player even if the skill between the two is actually the same
|
On September 05 2010 05:14 leve15 wrote:Show nested quote +On September 05 2010 04:52 rackdude wrote:On September 04 2010 23:24 Shorack wrote:
Since you know the entire population, there is no need at all for statistics of significance.
. Absolutely not true. What could be happening in the population could be random chance. If you created a population randomly with balanced races, you won't get 33% 33% 33% (forget random to make it easier), you get something slightly off. In the long run you would. After thousands of games, would you call the differences in that graph slight? Use your noggin.
No, I wouldn't call those slight. That's why the statistics on those graphs came out significant. Use your noggin and please read before you make snarky comments.
(I like how you cut my quote off before it got to statistical significance. Please don't misquote people and please read their whole post. Thank you.)
|
On September 05 2010 03:52 Scarmath wrote:Show nested quote +I'll try to repeat myself in an attempt to make it clearer compared to my first attempt: there is no question of significance, every difference is significant. There is only a question of relevance (are the differences we see big enough to act upon?) and for the relevant ones: what's causing it. I think, perhaps, you don't understand exactly. Let me state all this explicitly. It is clear that the proportions of the top 100 players are not equal to the proportions of the top 1000. The question is, how likely is that to be the result of random variation? We're not measuring sampling error. We don't need to measure sampling error. We do need to measure confidence. That is the point of both the binomial calculation and the chi-square test. We already know there is a difference, but we need to test how statistically significant that difference is. In both cases, we can say with 95% certainty that the proportions of the top 100 players are not different due to random chance. That is, the deviation is statistically significant. These are steps we need to take even when we have the entire population.
I see your point and i understand that we want to know whether it's some odd occurrence that just happens when dealing with something as complex and odd as human behavior. However, i don't agree with your method.
I disagree on two levels:
1. you are leaving out information we have. (this reduces the power and as a result prevents the drop of H0 where it would be dropped using full information)
2. i believe you might be forgetting to which random factor insignificance actually refers here.
1. i can only keep repeating this till either you agree or you give an explanation for not doing it: finite sample correction*. the exact proportions for the population are known. we can say with 100% certainty that the proportions aren't 30/30/30/10, but ab/cd/ef/gh (whatever they are, probably mentioned somewhere in the thread)
2. in case of insignificance, it means that if the population fits the H0, we'd still get the result found too often if we performed the same test on a new sample from the same population. So it refers to the randomness involved in selecting subjects for the sample. But here, the subjects observed equal the whole population. So there is zero chance of having such random effects, because we're not taking only part of the population and leaving a part out. (by which it's possible that just due to random chance, we get a sample that's too unrepresentative for the whole population and hence has different values from the population) And it's exactly that, that you're testing for when performing something of the likes of a binomial test: you're using techniques that are meant to cope with the limitations of a sample and you're using them on a full population.
Again, i totally agree that it's important to find out whether the differences are due to a factor of importance and whether they're large enough to act upon, but statistics is not meant for this**. (well, there are some tests for relevancy, but the one or two i've encountered did nothing but give you another number upon which you had to decide if it was relevant or not :p)
*Just to make sure, do you know what i mean when i write: root((N-n)/(N-1))
with: N=population size, n=sample size.
**Maybe it becomes clear by giving you the totally opposite situation of the one you refer to: say you have a huge sample out of an infinite population. Even very small differences tend to become significant then. Say happiness of men vs women at the workplace. You can find on a stapelschaal that women score 7.3 and men 7.18 and that difference being significant. Are you going to report that to management? I hope not, the difference is so small that it's not worth any efforts. It's not because i say that all those differences are significant that i mean they're all relevant nor am i claiming that they can't be the result of the complexities ('randomness') of human behavior.
On September 05 2010 04:52 rackdude wrote:Show nested quote +On September 04 2010 23:24 Shorack wrote:
Since you know the entire population, there is no need at all for statistics of significance.
. Absolutely not true. What could be happening in the population could be random chance. If you created a population randomly with balanced races, you won't get 33% 33% 33% (forget random to make it easier), you get something slightly off. Statistical significance tells you "if the races were balanced, there would be a P percent chance of this happening". When you get a number like .01%, you go "wow, there is almost no chance this randomly occurred". However, if you look at the data, you can make an inference from something that is random. For example, if you flipped a coin 3 times, and you saw all heads, if you looked at the graph you'd go "wow, this is definitely heads biased." Statistics would tell you though "hey dude, chill. There is a 12.5% chance of that happening randomly, so I wouldn't jump to conclusions just yet". That is why we are using statistics.
You can't create random populations. The population needs to confirm to the research question. You can create random samples though. As your post is now, i disagree (assuming we can achieve perfect balance in the broad sense (appeal), which is ofc not possible, so just as a thought-experiment.) Replace population with sample in your post and i'll completely agree.
|
On September 05 2010 08:03 Shorack wrote:Show nested quote +On September 05 2010 04:52 rackdude wrote:On September 04 2010 23:24 Shorack wrote:
Since you know the entire population, there is no need at all for statistics of significance.
. Absolutely not true. What could be happening in the population could be random chance. If you created a population randomly with balanced races, you won't get 33% 33% 33% (forget random to make it easier), you get something slightly off. Statistical significance tells you "if the races were balanced, there would be a P percent chance of this happening". When you get a number like .01%, you go "wow, there is almost no chance this randomly occurred". However, if you look at the data, you can make an inference from something that is random. For example, if you flipped a coin 3 times, and you saw all heads, if you looked at the graph you'd go "wow, this is definitely heads biased." Statistics would tell you though "hey dude, chill. There is a 12.5% chance of that happening randomly, so I wouldn't jump to conclusions just yet". That is why we are using statistics. You can't create random populations. The population needs to confirm to the research question. You can create random samples though. As your post is now, i disagree (assuming we can achieve perfect balance in the broad sense (appeal), which is ofc not possible, so just as a thought-experiment.) Replace population with sample in your post and i'll completely agree.
You are right, but it actually depends completely upon where your model starts. For most models, experiments, scientific papers, etc, you are completely correct. The population is what is and the sample is what is measured. But that's because the ideas dealing with "random populations" are already dealt with in the mathematics.
An example is like this. Participants enter a room where there is Card A and Card B. Assume there is no preference for either card. Participants pick a card and are now designated as group A or B. From this you create a field of theoretically possible populations from the different combinations of card picking that is possible. From this theoretical model, you can ask the question, "if I were to randomly pick a population, what is the chance I pick one that matches the population that I measured?". This is what I mean by "create a random population", it's like theoretically picking a card from your hand of possibilities. I probably should have said "take an arbitrary p element of the set of possible populations", and I probably shouldn't have said you won't get 33% 33% 33% because there could exist at least one population with that distribution. But I think you get the point.
Good call because you cannot take a random population in any empirical science because the population is defined as what exists. But I was speaking from a mathematical standpoint that wasn't measuring what exists, but rather the probability of such a population existing given the model we have created (which is what the simplified formulas in non-upper division statistics classes give you). I guess it's a slip we make these days since with computers we actually do "create" random populations for models, though we should be saying we are taking a possible random population.
|
I disagree on two levels:
1. you are leaving out information we have. (this reduces the power and as a result prevents the drop of H0 where it would be dropped using full information)
2. i believe you might be forgetting to which random factor insignificance actually refers here.
1. i can only keep repeating this till either you agree or you give an explanation for not doing it: finite sample correction*. the exact proportions for the population are known. we can say with 100% certainty that the proportions aren't 30/30/30/10, but ab/cd/ef/gh (whatever they are, probably mentioned somewhere in the thread)
2. in case of insignificance, it means that if the population fits the H0, we'd still get the result found too often if we performed the same test on a new sample from the same population. So it refers to the randomness involved in selecting subjects for the sample. But here, the subjects observed equal the whole population. So there is zero chance of having such random effects, because we're not taking only part of the population and leaving a part out. (by which it's possible that just due to random chance, we get a sample that's too unrepresentative for the whole population and hence has different values from the population) And it's exactly that, that you're testing for when performing something of the likes of a binomial test: you're using techniques that are meant to cope with the limitations of a sample and you're using them on a full population.
1) 30P/30T/30Z/10R is not the expected distribution. Racial preference is not dependent on balance, so there is no reason to expect that a perfectly balanced game would have an even distribution of race preference. The actual proportions for the entire top 1000 are ~36P/35T/23Z/6R. What I am testing is if each individual bin of 50 matches those proportions. That is, is the proportion of racial distribution independent of rank.
The null hypothesis of my test is that the bin being tested matches that distribution.
In the top 2 bins, in all the tests I've run, this null hypothesis has always been rejected.
2) What is happening in the population COULD be the result of random chance. Say a player has his power go out, or suffers from lag. These are not "balance" related issues, but the could result in undeserved losses. How can we judge if these factors shaped the proportions of the top tier players? We can use statistical methods to measure how unusual these proportions are. The chi-square test I used is specifically intended to measure how closely observations (that is, the actual preference of players) match an expected proportion (that is, the proportions of the entire population under consideration).
I hope this clears up some of your concerns.
|
On September 05 2010 04:40 Scarmath wrote:I've uploaded a version of the spreadsheet I've been working on here: http://www.mediafire.com/?uta56tbvr5w8ectThe easiest way to play with it is by replacing the Raw table with other information copy and pasted from SC2Ranks.com. (Copy and paste it from Firefox. Chrome doesn't copy the Alt-test I use to identify race, and I haven't tested it in Internet Explorer). The rest of the sheets should update automatically. Other things may be messed with, but may require more extensive fiddling to work. Still working on this a few hours at a time.
ty! I was looking for this :D
|
Ok guys,CLEARLY random needs a buff.
The reason zerg is low and terran is high is because there are 3x more terran. Is random the weakest choice because its lowest in the %? No.
Also, Protoss is favored in tournys now.
|
On September 06 2010 05:57 ibreakurface wrote:
The reason zerg is low and terran is high is because there are 3x more terran. Is random the weakest choice because its lowest in the %? No.
If you read the thread, you'll see that a lot of the posts are about how the Terran point distribution differs from what you'd expect based on the number of players per race.
|
Match up win percentages would be a lot more helpful to be honest.
I am not going to get into my thoughts on balance or lack thereof, but these numbers indicate race popularity at various levels. Rule number one of trying to find a causing agent of something is this 'Correlation does not mean causation'.
|
Match up win percentages would be a lot more helpful to be honest. And are not here right now, there are some projects ongoing to see the mu win percentages but it will take some time.
|
I updated SC2Ranks to do the column style mappings instead of the stacked percentages: http://sc2ranks.com/stats/race/all/1
I'm going to try and get some data available in CVS format too so you can do the stats without having to go through some elaborate process to get the data off of the site.
|
On September 05 2010 09:38 rackdude wrote:Show nested quote +On September 05 2010 08:03 Shorack wrote:On September 05 2010 04:52 rackdude wrote:On September 04 2010 23:24 Shorack wrote:
Since you know the entire population, there is no need at all for statistics of significance.
. Absolutely not true. What could be happening in the population could be random chance. If you created a population randomly with balanced races, you won't get 33% 33% 33% (forget random to make it easier), you get something slightly off. Statistical significance tells you "if the races were balanced, there would be a P percent chance of this happening". When you get a number like .01%, you go "wow, there is almost no chance this randomly occurred". However, if you look at the data, you can make an inference from something that is random. For example, if you flipped a coin 3 times, and you saw all heads, if you looked at the graph you'd go "wow, this is definitely heads biased." Statistics would tell you though "hey dude, chill. There is a 12.5% chance of that happening randomly, so I wouldn't jump to conclusions just yet". That is why we are using statistics. You can't create random populations. The population needs to confirm to the research question. You can create random samples though. As your post is now, i disagree (assuming we can achieve perfect balance in the broad sense (appeal), which is ofc not possible, so just as a thought-experiment.) Replace population with sample in your post and i'll completely agree. You are right, but it actually depends completely upon where your model starts. For most models, experiments, scientific papers, etc, you are completely correct. The population is what is and the sample is what is measured. But that's because the ideas dealing with "random populations" are already dealt with in the mathematics. An example is like this. Participants enter a room where there is Card A and Card B. Assume there is no preference for either card. Participants pick a card and are now designated as group A or B. From this you create a field of theoretically possible populations from the different combinations of card picking that is possible. From this theoretical model, you can ask the question, "if I were to randomly pick a population, what is the chance I pick one that matches the population that I measured?". This is what I mean by "create a random population", it's like theoretically picking a card from your hand of possibilities. I probably should have said "take an arbitrary p element of the set of possible populations", and I probably shouldn't have said you won't get 33% 33% 33% because there could exist at least one population with that distribution. But I think you get the point. Good call because you cannot take a random population in any empirical science because the population is defined as what exists. But I was speaking from a mathematical standpoint that wasn't measuring what exists, but rather the probability of such a population existing given the model we have created (which is what the simplified formulas in non-upper division statistics classes give you). I guess it's a slip we make these days since with computers we actually do "create" random populations for models, though we should be saying we are taking a possible random population.
First of all, interesting post.
Second, i'll try to implement my remarks to scarmath in this same post.
Generating random populations does indeed happen. (i'm mainly thinking of the bootstrapping procedure for predictive modeling, e.g. churn models)
But there, the point is that we're interested in an unknown future population and we want to make sure the model will be robust enough for that.
Here, we're interested in the current population and in our case, it's fully known. The goal is not prediction.
Then there is the use of Scarmath's bins.
Using those bins, i see the point of chi-square, since you are comparing two different populations. (even if the distribtuion of races is the same, they're still two different populations by their definition)
This may sound like semantics, but as i understood it earlier, i believed you were putting forward a certain race distribution and then used the chi-square test to see if the diamond 600+ (or whatever+, i'm not arguing here about exact numbers, but about the method) is a different population or not. (in that case, if they weren't, that would be arguing that the actual population could very well be that proposed population, which would be nonsense.)
Just to indicate that last point, the correct formula for the standarderror in the binomial case would be root(p*(p-1)/n)*root((N-n)/(N-1)) With N=population size, n=sample size.
Since the 'sample' is the actual population, N=n and so the standarderror becomes 0.
I still have some doubts about the bin approach, but since i can't base them for myself on some statistical foundation and i don't want to be irrational and obstructing at the same time, so keeping Wittgenstein's famous saying in mind (Wovon man nicht sprechen kann, darüber muss man schweigen), i won't try getting in on that.
|
Is anyone taking into account how much each race plays? If for whatever reason people who play terran tend to play more, then it would be inflated as well. What are the number of games per race at these levels?
If you see higher numbers of terran games at that level, then it makes perfect sense that there would be more. If you see terran with about the same or less games played than the others, then it might suggest an imbalance at the higher levels.
Also, it will be interesting to see what will happen come 1.1 when tanks get reduced a bit. I'm a terran player and IM sick of seeing tanks :-p
|
It's true there's a lot of problems with this model of race selection. Nonetheless this is very cool, and I hope for Orb to come in here and declare it mathematically proven that terran is imba.
|
Correct me if I'm wrong, but let's say theres a difference in the amount of players who play each race (as it is right now)... just for example: 40% Terran, 40% Protoss and 20% Zerg. Now, if the game were perfectly balanced and every player played a nearly perfect (humanly possible) game, the top 10/top100/top1000 should also have about 40% Terran, 40% Protoss and 20% Zerg, right? So any lower/higher number should point to imbalances (at least in the metagame if not in the game itself), is that correct?
edit: and what kind of effect on the top x players distribution would it have if only one of the 3 interracial matchups was imbalanced and the others were perfectly balanced? lets say Protoss vs Terran would be 90% in favour of Terran and ZvT as well as ZvP was 50/50 either way. How would that change the distribution in an AMM ladder which (theoretically) almost perfectly assigns players according to their internal rating?
|
Just a question, and I'm sorry if I missed someone else asking, but have you guys doing statistical analysis of this data tried simply disregarding the random players? I mean, as far as I could tell from your different tests and such(I'm not particularly familiar with them) the random players only really seemed to make things a lot more complicated.
On the basis that a significant amount of random players(more than maybe 2-3) would almost certainly play as 1/3 terran, 1/3 zerg, 1/3 protoss wouldn't simply ignoring the random players completely make your tests much more revealing? You wouldn't get the weird dips because a few random players were clustered around a certain rating.
I know disregarding data is a big nono in general, but what purpose does examining random players really serve? They play random, after all?
|
heishe, the problem is that there are no samples of games that have been played perfectly, meaning that there will always be a difference from the 'ideal' stats and what really shows.
Furthermore, as I said before, you should look at the number of games each of these players plays.
Top points are (i'll go top 20):
01 - 1823 - Toss
02 - 1780 - Zerg
03 - 1750 - Toss
04 - 1749 - Terran
05 - 1706 - Toss
06 - 1684 - Terran
07 - 1677 - Terran
08 - 1666 - Toss
09 - 1665 - Toss
10 - 1660 - Terran
11 - 1654 - Random
12 - 1652 - Zerg
13 - 1652 - Toss
14 - 1651 - Toss
15 - 1646 - Toss
16 - 1644 - Terran
17 - 1644 - Terran
18 - 1633 - Terran
19 - 1633 - Terran
20 - 1628 - Terran
If you look at that, almost all the top 20 toss are ABOVE terran. The only ones not really represented about equally are zerg, which everyone agrees has problems (I think its because they lost things like the "dark swarm" which shields zerg's 'mostly' melee / close range units until they could get to a good fighting position. I think if they brought that back it would really balance thigns out.
But you can't just look at the points. you have to look at the games played too, and the win rations of them. If #1 played 200 more games than #5, then it makes sense that his points are higher.
If you take the points + ratio, pretty much the top 10 are 2 toss and the rest terran.
If you take just ratio, then there are 2 terran in the top 10 and the rest are zerg and toss.
Another thing you have to look at is what type of match up it is. were they all evenly matched? was one favored over the other? If you lose but you're not favored, then even if you lose 50 games, you're points will not go down as fast as if you lose 10 where you are favored.
This is why there are so many statistics in the world, we've been analyzing them for years, yet nobody can ever seem to predict how they will end up. It's not really valid to take out a single stat and determine if a race is op, because you have to have a global view of the overall picture. It takes more than one chart.
If you look at games played by the top 10 (based off points) (i dont feel like typing top 20 numbers) it is:
01 - 425 - Toss
02 - 819 - Zerg
03 - 1077 - Toss
04 - 499 - Terran
05 - 286 - Toss
06 - 274 - Terran
07 - 224 - Terran
08 - 594 - Toss
09 - 546 - Toss
10 - 196 - Terran
To me the first noticeable number is the zerg. He's number 2, but he had to play a lot of games to get there. Also, #3 toss is the same way. The lowest average games played while still being in the top 10 is Terran with an average of about 298 games. then the poor lonely zerg with more than 800, and then toss with an average of 585 games.
If you look at the games played, with the ratio, with the points together, i think it suggests that the Terran ARE a little better represented than others. They average a full 250-300 games less than toss, and about 500 games less than zerg, yet are remaining decent contenders with them (again looking at the top 10 because i didnt want to go down to 20, although it looks like in general 10-20 terran seem to have between 200-500 games, so it might raise the terran average to mid 300's, while toss might go down a bit with the remaining toss in the top 20.)
Once again, I think if we saw a comeback of the "swarm", the zerg wins would dramatically increase. Honestly, it seems like toss and terran are ALMOST even, maybe with the smallist imba in favor of terran. Zerg have a veerrry poor showing.
Once again that is to be a little expected considering how few people play zerg v. terran and toss, but Even with that There should be more of a zerg showing than only 2 in the top 20.
Another thing to take into consideration is that there are more toss in diamond than terran, and fewer zerg. That puts a little extra weight on the terran showing in the top 20, a little less on the toss showing, and much more weight on zerg showings. I think coming up with some sort of weighted system that takes all of this into account would be the best mechanism to actually determine how each race should be performing, based on games played (which should also heavily weight the results considering more games typically = higher points), win/loss ratios, etc...
sorry if i rambled.... there was a lot i wanted to get out there
|
|
|
|
|
|
EPT![[image loading]](http://img839.imageshack.us/img839/7462/racediamond2.png)
