|
Win % in Diamond (accounting for player skill)
49.6% win rate for Protoss when fighting Terran.
52.8% win rate for Protoss when fighting Zerg.
49.6% win rate for Terran when fighting Zerg.
Now I'm hoping we could have a brief discussion of what that actually means.
I'm reading it as based upon hidden elo (skill) what percentage of games are won against different races.
So these stats only show the difficulty of the match-ups RELATIVE to each other.
The only reason each race doesn't add up to 100% is because there are not an equal number of players playing each race.
In other words these stats show that ZvP is harder for diamond zergs than ZvT.
But they do not necessarily show that P > Z.
Nor do they necessarily show that Z > T.
Because these stats cannot show what elo(skill) a player would reach if he changed races, since this is obviously pretty impossible.
Thoughts? Am I wrong?
|
+/- 3% means matchmaking is doing its job and that these statistics are pointless?
|
They show what I suspected, that PvZ is the most imbalanced matchup right now.
|
On October 08 2010 07:31 Fa1nT wrote:
+/- 3% means matchmaking is doing its job and that these statistics are pointless?
Well I don't think they're pointless.
But I think they only show what matchups are more difficult for each race and don't give any indication of "true" balance.
|
Is it just me or are these statistics meaningless for many reasons
|
On October 08 2010 07:31 Fa1nT wrote:
+/- 3% means matchmaking is doing its job and that these statistics are pointless?
Exactly. I can't believe people still don't realize this.
The matchmaking is designed so that when you lose, you play people that are worse, and when you win, you play people that are better....every game. This basically makes your record pan out to 50% no matter what race and matchup you're playing. Add to that the extreme range of skill found in diamond league (from "remind me wat a build order is again?" to top players in the country) and you get a completely meaningless set of statistics
|
Yup. Everyone else has said it perfect. Stats are designed to show match making works, that's about it.
|
8748 Posts
Let's take a look at win percentages on a race-versus-race basis. This is something else we look at to see how matchups are faring over many games. These numbers take individual player skill into account, which helps to avoid the 50% win/loss percentage effect that the matchmaking system can impart on straight win/loss ratios.
Blizzard claims that they're able, to some extent, to avoid that pitfall.
|
On October 08 2010 03:46 Sakarabu wrote:Seems there are alot of illiterate people in this thread. The stats take the ladder system into accountFrom the blog: Show nested quote +Let's take a look at win percentages on a race-versus-race basis. This is something else we look at to see how matchups are faring over many games. These numbers take individual player skill into account, which helps to avoid the 50% win/loss percentage effect that the matchmaking system can impart on straight win/loss ratios. Yes, they probably won't be 100% accurate, and yeah they are kind of useless, since the percentage of new players picking Terran in the lower leagues and losing is probably higher than any other race (Since this is the Terran expansion). The list of problems with these stats goes on.. But the arguement that these stats don't take into account the 50% win ratio ladder balancing has already been accounted for, so stop bringing it up.
this guy got it right.
edit: in other words the stats here matter (in terms of matchmaking has been taken account for). Tyler also got it right. Of course, I agree these stats aren't 100% accurate. It's not possible to get exact stats.
|
I want to know how they are "accounting for player skill" when coming up with these numbers. Not really sure how you go about doing that.
|
8748 Posts
On October 08 2010 07:41 nemahsys wrote:
I want to know how they are "accounting for player skill" when coming up with these numbers. Not really sure how you go about doing that.
That's essentially the question that is on the table already, the whole point of this thread.
|
Someone needs to go into Blizzard, and kick everyone on the balance team out. If these are the numbers they use to balance the game, it makes perfect sense why they keep nerfing protoss.
I think TL needs to make it's own balance notes and force them down blizzards throat. What are they doing?
|
• We're increasing roach range. This will allow roaches to be more effective in large groups, giving the zerg more options in the mid to end game.
• Fungal Growth will now prevent Blink, which will give zerg a way to stop endlessly Blinking stalkers which can be very challenging to deal with in large numbers.
• The Barracks are going to require a Supply Depot, which will impact a lot of early terran reaper pushes.
• The reaper speed upgrade will require the Factory, which is meant to weaken a lot of the early terran reaper attacks that dominate so many matches, especially in team games.
• We're making a number of increases to the health of zerg buildings, which will make the very vulnerable zerg technology structures more resistant to raids. We don't expect these hit point changes to have a super significant impact on the game, but the current numbers felt way too low.
Gosh blizz thank you for letting me go 14 hatch every game against T.
|
On October 08 2010 07:41 nemahsys wrote:
I want to know how they are "accounting for player skill" when coming up with these numbers. Not really sure how you go about doing that.
I think they disregard any gimme matches where one player has significantly higher MMR than the opponent.
|
On October 08 2010 07:41 nemahsys wrote:
I want to know how they are "accounting for player skill" when coming up with these numbers. Not really sure how you go about doing that.
The people at Blizzard are probably smart enough to create an algorithm to do this. I hope so at least  .
|
It'd be interesting for player skill considering their metric for player skill is what they use to generate their matchmaking, it'd seem that if a super imbalanced race existed and someone won all the time with that, then they would appear "skillful" regardless of actual "skill". The fact that "skill" is actually a fairly subjective term alone makes it seem hard to compensate for it.
|
My problem with terran isnt much at a high lvl, its that any idiot can MMM and get to diamond and its much much harder to counter skill wise as protoss. I can just assume every game they are going MMM, be correct in that assumption, and still not have more than a 60% chance at winning, meanwhile if anyone else knows EXACTLY what a protoss and zerg is going from second 1, they have a very very very good shot at stopping it.
These are of course my experiences at my skill level (low level diamond)
My biggest complaint really is that as protoss, you CANT rush a terran period if they have half a brain, this doesn't mean i want to be able to GG them every time in the first 5 minutes, but there needs to be a THREAT of it so they cant just sim city away and not have to worry about it for 10 minutes.... Meanwhile i need to scout constantly to check for cheese and if they are doing certain timing pushs.
I still enjoy protoss of course because i enjoy FF micro and all the various builds i have open to me, but it's fairly irritating that i can go from playing just protoss for 100 games into diamond, and then be able to play terran for the 3rd time ever and be able to move up in ranks in diamond, while i have been stuck with protoss for a week
|
On October 08 2010 07:49 Cade)Flayer wrote:Show nested quote +On October 08 2010 07:41 nemahsys wrote:
I want to know how they are "accounting for player skill" when coming up with these numbers. Not really sure how you go about doing that. I think they disregard any gimme matches where one player has significantly higher MMR than the opponent.
Is that just an assumption?
|
On October 08 2010 07:53 QueueQueue wrote:Show nested quote +On October 08 2010 07:49 Cade)Flayer wrote:On October 08 2010 07:41 nemahsys wrote:
I want to know how they are "accounting for player skill" when coming up with these numbers. Not really sure how you go about doing that. I think they disregard any gimme matches where one player has significantly higher MMR than the opponent. Is that just an assumption?
Yes, just trying to guess what Blizz are doing to get those stats.
|
On October 08 2010 07:44 Liquid`Tyler wrote:Show nested quote +On October 08 2010 07:41 nemahsys wrote:
I want to know how they are "accounting for player skill" when coming up with these numbers. Not really sure how you go about doing that. That's essentially the question that is on the table already, the whole point of this thread.
Yeah these stats are largely useless. It's like if i went to sc2ranks, grabbed the top 10, then stated the following facts:
In high-diamond, there racial ratios are:
10% random
20% zerg
20% protoss
50% terran
None of the above is incorrect, but completely useless and misleading if you don't know the context of how the stats were generated.
Thoughts on a pro-mod managed by TL (not released on 1. april)? I believe it's possible to make custom maps that store stuff between games, so a custom ladder could be possible? Then we'd have our own god damn stats.
|
They could assume that by skill, players are evenly distributed across the races. Then they look at matches between say a zerg that's more skilled than 95% of zergs vs a terran that's more skilled than 95% of terrans. Throw in more data points and interpolation and you might be on to something.
|
On October 08 2010 07:52 Varth wrote:
My problem with terran isnt much at a high lvl, its that any idiot can MMM and get to diamond and its much much harder to counter skill wise as protoss. I can just assume every game they are going MMM, be correct in that assumption, and still not have more than a 60% chance at winning, meanwhile if anyone else knows EXACTLY what a protoss and zerg is going from second 1, they have a very very very good shot at stopping it.
These are of course my experiences at my skill level (low level diamond)
My biggest complaint really is that as protoss, you CANT rush a terran period if they have half a brain, this doesn't mean i want to be able to GG them every time in the first 5 minutes, but there needs to be a THREAT of it so they cant just sim city away and not have to worry about it for 10 minutes.... Meanwhile i need to scout constantly to check for cheese and if they are doing certain timing pushs.
I still enjoy protoss of course because i enjoy FF micro and all the various builds i have open to me, but it's fairly irritating that i can go from playing just protoss for 100 games into diamond, and then be able to play terran for the 3rd time ever and be able to move up in ranks in diamond, while i have been stuck with protoss for a week
Learn to use force field properly. Then lets talk about this issue again.
|
On the Blizzard forums this is what the response to player skill was.
Show nested quote +"Doesn't race played throw it off though? You could never compare the skill of a terran and a protoss if the races weren't balanced. Say marines did 10 damage a shot, thus you would argue the terran has more skill?" "The system does account for this to an extent. After analyzing thousands of games and compiling trends based on player skill and race balance, our metrics will attempt to predict who the winner in a given match should be based on player skill and what it thinks is the favored race. These calculations will "correct" themselves over time as trends change. We do welcome feedback about the announced changes, but there are a considerable number of people outright disregarding the win/loss percentages we provided. Saying we conjured up these numbers to support our own ends, or that our numbers must be incredibly flawed, doesn't really add much to the discussion. The math was developed to disregard the way in which the matchmaking system attempts to keep players as close to a 50% win/loss ratio as possible. As stated in the blog, it's certainly not the only way we analyze balance, but it's very useful to us.
It does not really answer the question, but it adds alittle more to this conversation. They seem to already be on the defensive about the numbers.
|
really want to see that roach range.
it can be a nice addition for zergs to play with , or the thing that negates terran any chance to harrass with hellion or reaper depending on what they do.
|
On October 08 2010 08:01 snakeyes wrote:On the Blizzard forums this is what the response to player skill was. Show nested quote +"Doesn't race played throw it off though? You could never compare the skill of a terran and a protoss if the races weren't balanced. Say marines did 10 damage a shot, thus you would argue the terran has more skill?" "The system does account for this to an extent. After analyzing thousands of games and compiling trends based on player skill and race balance, our metrics will attempt to predict who the winner in a given match should be based on player skill and what it thinks is the favored race. These calculations will "correct" themselves over time as trends change. We do welcome feedback about the announced changes, but there are a considerable number of people outright disregarding the win/loss percentages we provided. Saying we conjured up these numbers to support our own ends, or that our numbers must be incredibly flawed, doesn't really add much to the discussion. The math was developed to disregard the way in which the matchmaking system attempts to keep players as close to a 50% win/loss ratio as possible. As stated in the blog, it's certainly not the only way we analyze balance, but it's very useful to us. It does not really answer the question, but it adds alittle more to this conversation. They seem to already be on the defensive about the numbers.
interesting, I have faith that these numbers are accurate enough to have some impact but not enough to determine balance.
|
I'd question these stats but the pretty much agree with what my opinion of the race matchups are.
Now looking at all of Diamond vs just the top few percentiles of Diamond will give very different results I'm sure. I would fully expect Protoss to look the strongest when looking at all of Diamond though.
|
i bet if zerg and toss had a win rate of 48% against terran, everyone would be jumping on the OP bandwagon.
|
On October 08 2010 07:59 Teddyman wrote:
They could assume that by skill, players are evenly distributed across the races. Then they look at matches between say a zerg that's more skilled than 95% of zergs vs a terran that's more skilled than 95% of terrans. Throw in more data points and interpolation and you might be on to something.
Well that would certainly be a very interesting statistic to view but I don't think that is what we are seeing.
Show nested quote +"Doesn't race played throw it off though? You could never compare the skill of a terran and a protoss if the races weren't balanced. Say marines did 10 damage a shot, thus you would argue the terran has more skill?" "The system does account for this to an extent. After analyzing thousands of games and compiling trends based on player skill and race balance, our metrics will attempt to predict who the winner in a given match should be based on player skill and what it thinks is the favored race. These calculations will "correct" themselves over time as trends change. We do welcome feedback about the announced changes, but there are a considerable number of people outright disregarding the win/loss percentages we provided. Saying we conjured up these numbers to support our own ends, or that our numbers must be incredibly flawed, doesn't really add much to the discussion. The math was developed to disregard the way in which the matchmaking system attempts to keep players as close to a 50% win/loss ratio as possible. As stated in the blog, it's certainly not the only way we analyze balance, but it's very useful to us.
Instead of all that nonsense, why don't they just tell us exactly what the stats are meant to mean?
|
On October 08 2010 08:01 snakeyes wrote:On the Blizzard forums this is what the response to player skill was. Show nested quote +"Doesn't race played throw it off though? You could never compare the skill of a terran and a protoss if the races weren't balanced. Say marines did 10 damage a shot, thus you would argue the terran has more skill?" "The system does account for this to an extent. After analyzing thousands of games and compiling trends based on player skill and race balance, our metrics will attempt to predict who the winner in a given match should be based on player skill and what it thinks is the favored race. These calculations will "correct" themselves over time as trends change. We do welcome feedback about the announced changes, but there are a considerable number of people outright disregarding the win/loss percentages we provided. Saying we conjured up these numbers to support our own ends, or that our numbers must be incredibly flawed, doesn't really add much to the discussion. The math was developed to disregard the way in which the matchmaking system attempts to keep players as close to a 50% win/loss ratio as possible. As stated in the blog, it's certainly not the only way we analyze balance, but it's very useful to us. It does not really answer the question, but it adds alittle more to this conversation. They seem to already be on the defensive about the numbers.
The thing is they're trying to point to these numbers for balance.
These numbers only prove that the game is "mostly balanced" which most reasonable people would agree on.
There are still huge glaring issues that cause dramatic differences like 1300 diamond terrans who build missile turrets vs zerg (I played this guy) and 1700 protoss players who have a devastating 4-gate, but don't know how to do anything else.
These will not be reflected in win/loss rates because their elementary strategies have them playing well above their real skill level due to singular, OP elements of the game.
|
3% is a lot. Think about how many millions of games have been played.
Edit:
Also agreed with above. Someone of less technical skill could perhaps be playing one race and because of the races potential imbalances would be tricking the system into thinking players are of equal skill.... right?
|
It's somewhat telling that out of all the numbers they could look at, they chose this.
While I am happy that they are trying to balance the game at all levels, and not just at the very top, it makes much more sense to me to try to create a level of parity at the top and then smooth out as you go down than to try to make piecemeal balance changes at the various levels then try to smash the final balance into shape at the top.
With so many possible factors at play it seems impossible to make generalizations based on a single metric and it's disappointing that they've disregarded so many and choose only a few as valid, especially when one like this seems to show a different picture than many others.
|
I would throw out that the big hits are the items that Blizz posted after the data:
I don't get the reaper change. Nitro boost requiring factory seems like a nerf to put itself into needlessness. Morrow's 5rax all in seems hurt, and at that timing woudln't it make more economical sense to just add Starport and drop?
Here are a few of the changes we currently have planned: •
We're increasing roach range. This will allow roaches to be more effective in large groups, giving the zerg more options in the mid to end game. •
Fungal Growth will now prevent Blink, which will give zerg a way to stop endlessly Blinking stalkers which can be very challenging to deal with in large numbers.
• The Barracks are going to require a Supply Depot, which will impact a lot of early terran reaper pushes.
• The reaper speed upgrade will require the Factory, which is meant to weaken a lot of the early terran reaper attacks that dominate so many matches, especially in team games.
• We're making a number of increases to the health of zerg buildings, which will make the very vulnerable zerg technology structures more resistant to raids. We don't expect these hit point changes to have a super significant impact on the game, but the current numbers felt way too low.
|
On October 08 2010 08:14 Jerubaal wrote:
With so many possible factors at play it seems impossible to make generalizations based on a single metric and it's disappointing that they've disregarded so many and choose only a few as valid, especially when one like this seems to show a different picture than many others.
The cynic in me says they published these stats because they wanted to make the game look really balanced to the uninformed.
|
On October 08 2010 07:34 SubtleArt wrote:Show nested quote +On October 08 2010 07:31 Fa1nT wrote:
+/- 3% means matchmaking is doing its job and that these statistics are pointless?
Exactly. I can't believe people still don't realize this. The matchmaking is designed so that when you lose, you play people that are worse, and when you win, you play people that are better....every game. This basically makes your record pan out to 50% no matter what race and matchup you're playing. Add to that the extreme range of skill found in diamond league (from "remind me wat a build order is again?" to top players in the country) and you get a completely meaningless set of statistics
What's worse is that BLIZZARD doesn't seem to fully realize this yet. They say they're "accounting for player skill" but they're measuring player skill based on the same ranking system they're using it to judge. They've mentioned they're aware of this pitfall and are secretly avoiding it, but I haven't seen any evidence that they're doing that correctly, and plenty of in-game experience to the contrary.
I constantly run into toss and terran players around 1k diamond who could never make it out of gold as zerg. It's these little experiences that show the imbalance. One guy didn't know what roaches are called -- they're "those vomity things." Another forgot brood lords existed. A great many of them have no idea what targets to focus on with their attacks... they'll go for the evo chamber or something and let my drones happily mine away right next to them. They'll banshee rush, get killed by mutas, and keep on building banshees because they don't know what to do when the rush fails. They almost never expand when they should. It's just so easy for P/T to get to this level by learning one extremely deadly unit composition per match-up (stalkers+colossus, etc) and a-moving their way to victory... they don't have to learn much about the game at all. Zerg can't get away with that.
|
I would have to agree, there is absolutely no way to determine an actual win-loss ratio that is meaningful because
1. There are 3 different races
2. The races are not the same
In most games where a rating system is in place, conditions are identical for both players (except for chess, which is why white wins 55% of the time). In starcraft, positions are not identical and therefore, player skill cannot be accurately determined. In chess, players play 50% of games as white, and 50% as black, so even though short term conditions are not the same, a rating is a calculation of both. Now unless there are players that can play all three races equally, a rating does not mean anything....however!
However, it is possible to determine matchup specific ratings.
For example, take the top 100 players from each race, and compare the average skill. Assuming that the talent is spread evenly amongst the groups, then it's possible to sorta extrapolate a little data, but with how little information there is, it's not likely to be very accurate.
|
imo blizzard should be aiming to balance the matchups only at the top levels rather then trying to balance the matchups at all levels sigh 
because once its balanced at the top, other/casual players will naturally gravitate towards the perceived easier races and it will balance itself out. i.e if zerg needs more skill to play than protoss, the zerg players who are less skilled are just gonna switch to protoss which will balance out the %.
|
Interesting that the racial stats are basically the same as in BW but completely reversed.
|
United States33388 Posts
Why is anyone talking about variance when blizzard has THE ENTIRE FUCKING SAMPLE OF EVERY GAME PLAYED?
|
I want to see the %'s for players who are exclusively Random.
|
On October 08 2010 07:57 Adeny wrote:
Yeah these stats are largely useless. It's like if i went to sc2ranks, grabbed the top 10, then stated the following facts:
None of the above is incorrect, but completely useless and misleading if you don't know the context of how the stats were generated.
The conceit of all of you who complain Blizzard can't make judgments without knowing how the stats are generated even though they are the ones who designed them are ridiculous.
Get over it. Blizzard knows the numbers better than you do. Whether or not they are making a good decisions on those numbers is open to debate.
|
Why is the thread about win percentages - is there a thread somewhere that I am not seeing that is talking about BARRACKS NOW REQUIRING A DEPOT and REAPER SPEED NOW REQUIRING A FACTORY? Those seem much more discussion-worthy.
|
On October 08 2010 08:14 Mutarisk wrote:
3% is a lot. Think about how many millions of games have been played.
3% doesn't become more just because it is extrapolated across millions of games. You're essentially saying "the small margin of error isn't small enough because the source is large as a whole," which makes very little sense.
Now if you want to argue that 3% is just too large in the first place thats a different discussion.
|
This patch is going to make me so giddy...I'm gonna proxy pylon some poor Terran's ramp to prevent barrack placement :D
|
On October 08 2010 08:36 Pyrrhuloxia wrote:
Why is the thread about win percentages - is there a thread somewhere that I am not seeing that is talking about BARRACKS NOW REQUIRING A DEPOT and REAPER SPEED NOW REQUIRING A FACTORY? Those seem much more discussion-worthy.
On the contrary, I find individual balance discussions pale in comparison to discussing the methods of balancing.
In fact, I would almost call it a META balance discussion, but I have been sternly rebuked by others far my superior and told to keep my impertinent suggestions to myself.
|
"The system does account for this to an extent. After analyzing thousands of games and compiling trends based on player skill and race balance, our metrics will attempt to predict who the winner in a given match should be based on player skill and what it thinks is the favored race. These calculations will "correct" themselves over time as trends change.
We do welcome feedback about the announced changes, but there are a considerable number of people outright disregarding the win/loss percentages we provided. Saying we conjured up these numbers to support our own ends, or that our numbers must be incredibly flawed, doesn't really add much to the discussion. The math was developed to disregard the way in which the matchmaking system attempts to keep players as close to a 50% win/loss ratio as possible. As stated in the blog, it's certainly not the only way we analyze balance, but it's very useful to us.
I would really like to know their algorithm. It would deserve a placement in a mathematics thesis if this was possible. With the algorithm the way it's designed it is IMPOSSIBLE to determine a win/loss ratio without equal conditions or somewhat equal conditions. Predicting who the winner is is based on ELO (or glicko-2 or trueskill), determining balance does not derive from this at all.
If only blizzard gave us access to a list of all games in an api then we could determine this ourselves.
Note:
Technically sc2 is a combination of 6 different games.
pvp
pvt
pvz
tvt
tvz
zvz
so this makes it even harder.
|
On October 08 2010 08:31 Waxangel wrote:
Why is anyone talking about variance when blizzard has THE ENTIRE FUCKING SAMPLE OF EVERY GAME PLAYED?
Because when you're calculating whether or not X race is going win against Y race, you're going to get a +/- variance in the model that can't be accounted for unless you literally include every possible variable in your model (ie. a meteor killing a player while he's playing the game). This is impossible, variance is present regardless of access to population data or not, because you're attempting to pigeonhole something that normally can't be explained via modeling into a nice and neat mathematical equation. You do the best you can, but it's not going to be 100% perfect.
|
Adeny, don't flame. TL is not flame ground.
-Don't quote a top-level foreigner to improve your flawed logic,
-pull a sample size of 10 (and one that isn't even close to real top 10),
-use %'s with 10 users,
-absolutes
-cussing
-flame incoherently...etc.
-digress from what OP wants
Quit being full of yourself and contribute to discussion or flame somewhere else. I doubt you actually feel a major imba problem to be frank.
==================================
3% is significant with the degree of information available. How the present the information is the problem.
*adjusted for player skill is quite questionable, but probably means if standardized for ELO of ____
example: [Race] V [Race] of 8,000 users for ELO of [Range].
They could have said okay Bronze ELO is 0-1000 or something Silver 1000-1300 etc. and let's pull 60% percentile of each division +/- 50 ELO to show a point. One can do so many things to "prove" a point in stats. We all are just frustrated at their deceptive behavior.
Being a Zerg, in ZvP I do feel I need to be a lot better to gain a small advantage. This is the only statistic that I feel is close to the current meta-game (the point of bringing up the stat). I still enjoy ZvP the most, though.
I think the PvT is the biggest problem, though, it says 50/50. Hmm.
|
It would be interesting to see the maths behind it, but it does make sense that if one race is overpowered (let's say Zerg for example) then the stats for Protoss would look bad against zerg, but good against Terran. It's even more complicated than that, but I do not think that we can use these stats to tell us anything truly meaningful about the game.
|
The problem is unavoidable.
The skill that blizzard has hidden or not still is a function of the the race being played and the players true skill.
Defining either of these two quantities or how they relate is too complicated to do and is also somewhat subjective.
I just dont see how stats that assume that a players rating reflects their skill can possibly be used to determine balance of the game.
Now If you could quantify those two traits then racial strength would depend on true skill and would be a complicated function that would incorporate how the balance between the races shift at as true skill varies and as experience with the race varies.
|
Note:
Technically sc2 is a combination of 6 different games.
pvp
pvt
pvz
tvt
tvz
zvz
so this makes it even harder.
Mirror match ups are not counted because they are balanced for either player.
|
On October 08 2010 08:39 VonLego wrote:Show nested quote +On October 08 2010 08:14 Mutarisk wrote:
3% is a lot. Think about how many millions of games have been played.
3% doesn't become more just because it is extrapolated across millions of games. You're essentially saying "the small margin of error isn't small enough because the source is large as a whole," which makes very little sense. Now if you want to argue that 3% is just too large in the first place thats a different discussion.
You are correct.
What I am arguing is that as the number of games increases, the number should become closer to zero difference which would represent "balance" (assuming they have an accurate measure of skill)
For the number of games that have been played, I think 3% is rather large.
|
On October 08 2010 08:35 mutantmagnet wrote:Show nested quote +On October 08 2010 07:57 Adeny wrote:
Yeah these stats are largely useless. It's like if i went to sc2ranks, grabbed the top 10, then stated the following facts:
None of the above is incorrect, but completely useless and misleading if you don't know the context of how the stats were generated.
The conceit of all of you who complain Blizzard can't make judgments without knowing how the stats are generated even though they are the ones who designed them are ridiculous. Get over it. Blizzard knows the numbers better than you do. Whether or not they are making a good decisions on those numbers is open to debate.
Wait are you trying to disagree with me? Blizzard would prefer to make themselves look good. You're assuming that a multi-billion dollar company with hundreds of thousands of stockholders would not try to shift the statistics in a way that would make them look better.
I'm well aware that Blizzard knows the numbers better than I do, THAT'S EXACTLY WHY THEY ARE USELESS. Only if they openly released the methods they used to arrive at those statistics would they be relevant in any way.
- Never does it state if the entire pool of players are used, for all you know it's only EU, only US etc.
- "Accouting for skill" could mean any number of things, but in essence it probably means "we tried every comparison and this is the one that makes us look best".
I'm not trying to claim I know more about Blizzards numbers than they do, but blinding trusting their very own statistics is extremely naive, especially with the lack of information.
|
On October 08 2010 07:47 KingAce wrote:
Someone needs to go into Blizzard, and kick everyone on the balance team out. If these are the numbers they use to balance the game, it makes perfect sense why they keep nerfing protoss.
I think TL needs to make it's own balance notes and force them down blizzards throat. What are they doing?
If you think Protoss deserves a buff, I laugh.
|
Can you even use this for anything that got to do with balance? If zerg is/were weaker then the other races and you start loseing games against the other races with about even skill lvl, then the system will just lower your ELO so you play players a bit weaker in skill then you and win rate will be about 50%.
Aint that how the system work?
|
On October 08 2010 08:48 StimCraft wrote:
Adeny, don't flame. TL is not flame ground.
-Don't quote a top-level foreigner to improve your flawed logic,
-pull a sample size of 10 (and one that isn't even close to real top 10),
-use %'s with 10 users,
-absolutes
-cussing
-flame incoherently...etc.
-digress from what OP wants
Quit being full of yourself and contribute to discussion or flame somewhere else. I doubt you actually feel a major imba problem to be frank.
WHAT?! Wait, are you sure you are reading my post? Cussing where? Flame incoherently, where? And digress? My post is very on point, let me break this down for you.
The point I was trying to prove was that any statistic is complete garbage if the method used to generate said statistics is released in full. I'm using a sample size of 10 and %'s with 10 users to show that you can make factual statistical statements and make them look exactly the way you want them to. Real top 10 or not it was just an example, but now I understand how Darwin must have felt. You're blindly assuming Blizzard is 100% correct, I'm demanding evidence to support their claims.
Again, I'm not "full of myself", I am rationally and logically explaining why the numbers actually mean nothing, so stop your higher-than-thou attitude, please.
|
What we're also aware of is that, while the numbers don't necessarily support the need for zerg changes across all leagues, the feedback from the community as well as our own play experience tells us that improvements are necessary to make zerg matchups feel and play better.
Regardless of how they compile the data for racial stats, they've shown the ability to ignore them in the same post they released them, evidenced by that quote and all 5 previewed changes being Zerg buffs.
|
On October 08 2010 07:29 Klive5ive wrote:
Win % in Diamond (accounting for player skill)
49.6% win rate for Protoss when fighting Terran.
52.8% win rate for Protoss when fighting Zerg.
49.6% win rate for Terran when fighting Zerg.
the number tells ZvT and PvT is fine
than, why BZ want to buff zerg ??
very strange
|
On October 08 2010 07:41 nemahsys wrote:
I want to know how they are "accounting for player skill" when coming up with these numbers. Not really sure how you go about doing that.
They could determine the true relative ranking of players by looking at just the results from mirror matches. Not hard at all to do.
|
it just means that their matchmaking system works. good zerg and protoss players are being placed against less skilled terrans, and it equals out.
the only statistics that indicate the imbalance are the very top ranked terrans, because there are almost no zergs good enough to match them (had to say almost no because you all would go WHAT ABOUT FRUITDEALER). if they looked at the very top 50 players from each region, i think they'd get a better sense of balance than "oh well our matchmaking system that is designed to place players against players they will lose 50% of their games to is telling us that they are, in fact, losing 50% of their games. game balanced, guys!"
|
So, the barracks will require depot just because they want to delay "reaper pushes". Why don't just increase the reaper build time, or make them 2 food if that's the problem?
This could affect things like early marauder pressure in tvp just because they want to change the reaper timings.
|
I would like to point out that if you think about it, 3% isn't that big of an advantage. In particular I'm going with the assumption that the number of games played is large and thus the error in this statistic isn't unreasonably large (<10%). But even in chess, the general consensus is that white has about a 4% advantage over black (see http://en.wikipedia.org/wiki/First-move_advantage_in_chess). Now I don't know about you, but when I play a game of chess, I don't necessarily stand up and yell "IMBALANCED!!!" whenever I happen to play black. So I would say that the 3% is the least troubling of the statistics. I am more curious as to why we hear so many Zergs complaining and yet the statistics don't seem to reflect this feeling very well...
|
On October 08 2010 07:40 Liquid`Tyler wrote:Show nested quote +Let's take a look at win percentages on a race-versus-race basis. This is something else we look at to see how matchups are faring over many games. These numbers take individual player skill into account, which helps to avoid the 50% win/loss percentage effect that the matchmaking system can impart on straight win/loss ratios. Blizzard claims that they're able, to some extent, to avoid that pitfall.
But it really doesn't because ultimately, a player is gonna go close to 50/50 regardless of what match up he's playing, unless he's some statistical deviant who excels at one match up but is absolutely terrible at another. The fact that the skill range in diamond is so massive pretty much ensures this. I don't think Blizzard realizes just how useless the divisions are at the moment.
|
I'm zerg and my win ratio is about 53% at about 1k points. I win maybe 75% ZvZ's, win about 50% of ZvP and win maybe 25% of my ZvT's.
|
On October 08 2010 09:56 TanukTanukTuk wrote:I would like to point out that if you think about it, 3% isn't that big of an advantage. In particular I'm going with the assumption that the number of games played is large and thus the error in this statistic isn't unreasonably large (<10%). But even in chess, the general consensus is that white has about a 4% advantage over black (see http://en.wikipedia.org/wiki/First-move_advantage_in_chess). Now I don't know about you, but when I play a game of chess, I don't necessarily stand up and yell "IMBALANCED!!!" whenever I happen to play black. So I would say that the 3% is the least troubling of the statistics. I am more curious as to why we hear so many Zergs complaining and yet the statistics don't seem to reflect this feeling very well...
In chess it actually does matter, and for grandmasters the color they play is pretty important. In long series the goal is often to play an opening that sets you up in a position for a draw if you're black.
|
Does this mean P>Z>T>P in sc2? Hows that for a change!
Eh, seriously though, I dont think these stats mean much as they account only NA, and considering ppl in diamond league are pretty skilled, such close numbers are sort of expected. It however may be an indicator that the balance isnt as bad ppl think.
I personally find the percentage of Zerg players (23%?) more interesting. I am glad Blizzard will balance changes primarily focusing on improving the zerg on next patch.
|
On October 08 2010 07:32 Pyrrhuloxia wrote:
They show what I suspected, that PvZ is the most imbalanced matchup right now.
\Then y do everyone cry about TvZ?
|
On October 08 2010 08:31 Waxangel wrote:
Why is anyone talking about variance when blizzard has THE ENTIRE FUCKING SAMPLE OF EVERY GAME PLAYED?
ahahaha this is awesome, made me lol
it's funny because it's so obvious when pointed out
|
On October 08 2010 10:08 HiHiByeBye wrote:Show nested quote +On October 08 2010 07:32 Pyrrhuloxia wrote:
They show what I suspected, that PvZ is the most imbalanced matchup right now. \Then y do everyone cry about TvZ?
Because T can easily win against Z up until mid-diamond with very very simple play. Each action on a low-mid level requires a much more difficult response from Z.
On high level, I believe Cool and Artosis both mentioned huge problems with PvZ - especially regarding the pylon+cannon block at the bottom of the Main ramp, and the huge economic advantage P can get with 2gate pressure.
|
2 Things I think Blizzard should mention:
1. Clarify what "accounting for player skill" means. I don't think we're getting the actual numbers here, I think we're getting modified versions.
2. How far back do the stats date? Launch? Terran were a lot worse at launch, and lots of newbies jumped on the terran bandwagon and brought the W/L ratio down. Not to mention, the number of terran builds and strategies has made them extremely diverse and hard to stop.
|
I agree with Ansat that the numbers Blizzard has provided is of no use for us readers. I can't iron out weather Blizzard does a good job balancing. Not with this! Either they don't know what they are doing or they are consciously dumbing down the presentation of the data to manipulate us :-S
Statements that statistics doesn't help in balancing gives me an uneducated impression. Logically you should be able to get a estimation of race balance through the game statistics Blizzard has at hand. Teddymans suggestion should be able to decide witch race need balancing.
The first thing Blizzard needs to address is the unpopularity of the zerg race, wich is a balance issue in itself. Making it a more attractive race to play should be a top priority.
Blizzard, please incubate a few staticians for further use in your nerdswarm!
|
The number of players playing each race is indicative of how difficult or interesting is this race.
Protoss are played 38.5% of the time.
Terran are played 38.0% of the time.
Zerg are played 23.5% of the time.
Now, apart from this, a good measure for comparison is the distribution of the players of each race on the different leagues and the average points of the players of each race in each league, instead of the win ratio.
|
On October 08 2010 10:17 Blueblister wrote:
I agree with Ansat that the numbers Blizzard has provided is of no use for us readers. I can't iron out weather Blizzard does a good job balancing. Not with this! Either they don't know what they are doing or they are consciously dumbing down the presentation of the data to manipulate us :-S
Statements that statistics doesn't help in balancing gives me an uneducated impression. Logically you should be able to get a estimation of race balance through the game statistics Blizzard has at hand. Teddymans suggestion should be able to decide witch race need balancing.
The first thing Blizzard needs to address is the unpopularity of the zerg race, wich is a balance issue in itself. Making it a more attractive race to play should be a top priority.
Blizzard, please incubate a few staticians for further use in your nerdswarm!
I don't think people as proficient at writing a match making system that works this well wouldn't be able to point out to their designer colleagues which factor are affecting the statistics in what why and how to filter it for reporting.
|
I dont see how people can say these stats are meaningless without also saying other stats like "4 hydras beat 3 marauders" is meaningless. These stats mean something. They mean that Zerg players are beating Terran players. Something I think most of us are learning to accept. Just a vocal minority left....
They mention in the article that they are considering ways of nerfing Protoss to alter Protoss vs Terran late game. Although I do not cry of imbalance, I am open to changes in Protoss vs Terran that favor T when the game reaches what I call "macro stages".
|
AHHHHHHHH, honestly I dunno what to make of the stats, because these stats include all the newbs that somehow made it into diamond via cheese strats. People that only play a specific race because they know a cheap BO for that race, and that throws off the % some. Such as, imagine if their was no cannon rush, protoss would be played the same amount as other races, and the game is still so young that lots of people who deserve to be diamond are stuck in platinum/gold.
Maybe im jus a idiot...If anything It says, your general zerg player has more talent than your general terran player, protoss somewhere in the middle.
|
Let's take a hypothetical case where everyone sucks equally - we would expect win ratios for all races to be about even. And they always ARE about even (roughly). And somehow I suspect that no matter what they do to the balance, the win/loss ratios will be about split among everyone, because in the grand scheme of things the wins and losses are a lot more chaotic than we think.
Balance based on the total dynamic of ALL players is going to be very difficult to see, you would have to be carefully watching percent changes of 0.001 percent or lower. For true balance, you need to look at the absolute top level of the playing field and make adjustments there. The effects will "trickle down" to the bottom of the ladder accordingly, where rushes and rage quits are the dominant effectors anyway.
I want to see changes that make the game more dynamic and result in more options. Like Pokemon, which decided to make like 30 different status variables all of which have different advantages or disadvantages over others.
A WILD HYDRALISK APPEARS!
COLOSSUS USES THERMAL LANCE!
IT'S SUPER-EFFECTIVE!
No nerfs I say! If one races fares poorly, give it more options! I yearn for the day when Toss plays something other than Zealot/Stalker/Colossus every game.
|
On October 08 2010 10:22 Techno wrote:
I dont see how people can say these stats are meaningless without also saying other stats like "4 hydras beat 3 marauders" is meaningless. These stats mean something. They mean that Zerg players are beating Terran players. Something I think most of us are learning to accept. Just a vocal minority left....
They mention in the article that they are considering ways of nerfing Protoss to alter Protoss vs Terran late game. Although I do not cry of imbalance, I am open to changes in Protoss vs Terran that favor T when the game reaches what I call "macro stages".
They are meaningless because you don't know wtf they portray, it could literally be anything. 4 hydras beating 3 marauders is also completely meaningless without additional information and I don't think you could find anyone to disagree with that.
|
On October 08 2010 10:02 Hypatio wrote:
I'm zerg and my win ratio is about 53% at about 1k points. I win maybe 75% ZvZ's, win about 50% of ZvP and win maybe 25% of my ZvT's.
You should work on your ZvT 
On October 08 2010 10:30 Adeny wrote:Show nested quote +On October 08 2010 10:22 Techno wrote:
I dont see how people can say these stats are meaningless without also saying other stats like "4 hydras beat 3 marauders" is meaningless. These stats mean something. They mean that Zerg players are beating Terran players. Something I think most of us are learning to accept. Just a vocal minority left....
They mention in the article that they are considering ways of nerfing Protoss to alter Protoss vs Terran late game. Although I do not cry of imbalance, I am open to changes in Protoss vs Terran that favor T when the game reaches what I call "macro stages". They are meaningless because you don't know wtf they portray, it could literally be anything. 4 hydras beating 3 marauders is also completely meaningless without additional information and I don't think you could find anyone to disagree with that.
They portray exactly what they say they portray. Race vs Race win/loss ratios at all the different leagues.
What kind of statistic DOES mean something to you?
To me this means that the game is pretty damn balanced, of course it isnt perfect, but its pretty damn balanced. How can anyone disagree? By disagree you would say that the game is very unbalanced, and to me that means that it is possible for a player to win 100% of his/her games by doing a particular thing which has no response. I'm sure players will say things like "MMM has no response", but that's getting really old and I will not waste my cyber breath on heathens such as that.
|
On October 08 2010 09:52 tianGO wrote:
So, the barracks will require depot just because they want to delay "reaper pushes". Why don't just increase the reaper build time, or make them 2 food if that's the problem?
This could affect things like early marauder pressure in tvp just because they want to change the reaper timings.
I believe on the Korean server there is a double rax before depot build going around, which I'm lead to believe is ridiculous in ZvT. This could be a response to that.
|
This stuff is pretty interesting.
Did blizzard ever say why their "hidden rating" is hidden?
|
PvZ makes sense
PvZ has pretty defined periods of strength that most protoss can learn quickly
Early game has 4 gate and void rays. Both have game ending potential but aren't too hard to stop once you play more ZvP. These will probably win you 60% PvZ at low low diamond. Then you move up a little and you have the colossus timing attacks.
4 gate still sort of works but I'm talking about 4gate 2 robo or something similar off 2 base. Very simple to learn and execute for your mid level players.
Late game requires more of a learning curve but only those 2 stages of the game are really needed to heavily influence the win percentage. So what I take from that is protoss vs zerg is easier to learn then zvp.
ZvT really is baffling. I don't even know what to say to those terrans in diamond that account for 51% loss.
|
Half of those games where protoss is concerned, are either a failed 4 gate, or a won 4gate. These statistics do not say much about game balance, but they do show that matchmaking is doing its job (to a decent extent) The mid/low diamond plays are absolutely horrid.
|
On October 08 2010 10:32 Techno wrote:Show nested quote +On October 08 2010 10:02 Hypatio wrote:
I'm zerg and my win ratio is about 53% at about 1k points. I win maybe 75% ZvZ's, win about 50% of ZvP and win maybe 25% of my ZvT's. You should work on your ZvT Show nested quote +On October 08 2010 10:30 Adeny wrote:On October 08 2010 10:22 Techno wrote:
I dont see how people can say these stats are meaningless without also saying other stats like "4 hydras beat 3 marauders" is meaningless. These stats mean something. They mean that Zerg players are beating Terran players. Something I think most of us are learning to accept. Just a vocal minority left....
They mention in the article that they are considering ways of nerfing Protoss to alter Protoss vs Terran late game. Although I do not cry of imbalance, I am open to changes in Protoss vs Terran that favor T when the game reaches what I call "macro stages". They are meaningless because you don't know wtf they portray, it could literally be anything. 4 hydras beating 3 marauders is also completely meaningless without additional information and I don't think you could find anyone to disagree with that. They portray exactly what they say they portray. Race vs Race win/loss ratios at all the different leagues. What kind of statistic DOES mean something to you? To me this means that the game is pretty damn balanced, of course it isnt perfect, but its pretty damn balanced. How can anyone disagree? By disagree you would say that the game is very unbalanced, and to me that means that it is possible for a player to win 100% of his/her games by doing a particular thing which has no response. I'm sure players will say things like "MMM has no response", but that's getting really old and I will not waste my cyber breath on heathens such as that.
The disagreeing part was about 4 hydras beat 3 maraduers being a useless stat.
So again, I'm NOT saying the game is imbalanced... What I am saying is that these stats are not worth much at all because there are variables invovled in their algorithm for generating those stats that we have NO idea what they are. Mainly what their "compensation" for play skill or whatever is, there's also the fact that we don't know if they include stats from the beta (it sounds ridiculous but Blizzard would obviously want to look good, so why not if they're telling the truth, right?), or if it includes the entirety of the stats from release or not.
Worthy stats would be stats that have every variable and how much weight they carry released at the VERY least, preferably the entire algorithm would be released, but chances are Blizzard cannot do that.
|
On October 08 2010 07:48 Crabman123 wrote:
• We're increasing roach range. This will allow roaches to be more effective in large groups, giving the zerg more options in the mid to end game.
• Fungal Growth will now prevent Blink, which will give zerg a way to stop endlessly Blinking stalkers which can be very challenging to deal with in large numbers.
• The Barracks are going to require a Supply Depot, which will impact a lot of early terran reaper pushes.
• The reaper speed upgrade will require the Factory, which is meant to weaken a lot of the early terran reaper attacks that dominate so many matches, especially in team games.
• We're making a number of increases to the health of zerg buildings, which will make the very vulnerable zerg technology structures more resistant to raids. We don't expect these hit point changes to have a super significant impact on the game, but the current numbers felt way too low.
Gosh blizz thank you for letting me go 14 hatch every game against T.
When I first read this, I was like "haha, made my day especially the supply one, best troll ever ^^"
Then I was like : Wtf, they are really thinking about doing this O_o
I mean, I play zerg, and what I want is that blizzard gives me more openings, not that they cut terran openings down to bring it to our level ><
|
On October 08 2010 10:48 DoomSpirit wrote:Show nested quote +On October 08 2010 07:48 Crabman123 wrote:
• We're increasing roach range. This will allow roaches to be more effective in large groups, giving the zerg more options in the mid to end game.
• Fungal Growth will now prevent Blink, which will give zerg a way to stop endlessly Blinking stalkers which can be very challenging to deal with in large numbers.
• The Barracks are going to require a Supply Depot, which will impact a lot of early terran reaper pushes.
• The reaper speed upgrade will require the Factory, which is meant to weaken a lot of the early terran reaper attacks that dominate so many matches, especially in team games.
• We're making a number of increases to the health of zerg buildings, which will make the very vulnerable zerg technology structures more resistant to raids. We don't expect these hit point changes to have a super significant impact on the game, but the current numbers felt way too low.
Gosh blizz thank you for letting me go 14 hatch every game against T. When I first read this, I was like "haha, made my day especially the supply one, best troll ever ^^" Then I was like : Wtf, they are really thinking about doing this O_o I mean, I play zerg, and what I want is that blizzard gives me more openings, not that they cut terran openings down to bring it to our level ><
yeah wtf... terrans will feel so lame now that i won't have to worry about getting proxyraxed or shit :/ i thouroughly enjoy outmicroing noob terrans looking for ezwin :D
|
ok so how does this supply depot before barracks change affect T vs 6 pool ling rush on a 2 player map with set bases?
|
Why dont they just watch how randomplayers perform in their matchups?
|
These data are very simple to interpret:
-Blizzard is "accounting for player skill." That means that they are using a basic regression analysis which everyone learns in every stats class that is worth a damn. After running the regression, they find the extent to which race is correlated with victory independent of player skill. Regression is how we can learn anything at all from statistics; I'm sure they did this correctly.
-For those interested in what "player skill" is, there have been a wonderful series of posts regarding the ladder system that discuss the nature of the "hidden rankings" that determine match-ups and are different than the displayed ratings.
-The key problem with the statistics has already been mentioned: the hidden player skill value is determined by your wins and losses. Thus, if a racial imbalance exists, Zerg players will lose more often to terrans and will appear to be less skillful. That means that there is correlation between the independent and dependent variables in the regression, which makes the conclusions of the regression suspect.
-So what have we learned? Absolutely nothing. They have corrected for player skill to the best of their ability. But how can we really be sure if one race is more powerful than the other or if more talented player gravitate towards one race?
Finally, for those of you paranoid about what this data means for future patches, etc., please remain calm. The Blizzard team is made up of professionals who know damn well that their data is flawed. They made it public as a PR thing to show that they care about balance. Of course, it didn't work because there's nothing nerds like better than to complain about how they're the victims of some big, strong athlete...err...I mean corporation, so people complain and complain anyways. The data doesn't mean anything. You know that, I know that, Blizzard knows that. What more is there to say?
|
On October 08 2010 10:32 Techno wrote:Show nested quote +On October 08 2010 10:02 Hypatio wrote:
I'm zerg and my win ratio is about 53% at about 1k points. I win maybe 75% ZvZ's, win about 50% of ZvP and win maybe 25% of my ZvT's. You should work on your ZvT Show nested quote +On October 08 2010 10:30 Adeny wrote:On October 08 2010 10:22 Techno wrote:
I dont see how people can say these stats are meaningless without also saying other stats like "4 hydras beat 3 marauders" is meaningless. These stats mean something. They mean that Zerg players are beating Terran players. Something I think most of us are learning to accept. Just a vocal minority left....
They mention in the article that they are considering ways of nerfing Protoss to alter Protoss vs Terran late game. Although I do not cry of imbalance, I am open to changes in Protoss vs Terran that favor T when the game reaches what I call "macro stages". They are meaningless because you don't know wtf they portray, it could literally be anything. 4 hydras beating 3 marauders is also completely meaningless without additional information and I don't think you could find anyone to disagree with that. They portray exactly what they say they portray. Race vs Race win/loss ratios at all the different leagues. What kind of statistic DOES mean something to you? To me this means that the game is pretty damn balanced, of course it isnt perfect, but its pretty damn balanced. How can anyone disagree? By disagree you would say that the game is very unbalanced, and to me that means that it is possible for a player to win 100% of his/her games by doing a particular thing which has no response. I'm sure players will say things like "MMM has no response", but that's getting really old and I will not waste my cyber breath on heathens such as that.
The stats only show that the Matchmaking system is working.
The only stat that blizzard collects is the match outcome i.e. wins and losses. Based on your wins and losses vs other players wins and losses, you get somewhat ranked, this is their hidden rating.
|
The first graph here is 100x more telling than any of Blizz's stats:
http://www.sc2ranks.com/stats/race/us/1/
The very clear correlation between increase in Diamond points and increase in Terran percentage means one thing: According to the match-making system, the arbitrary Terran player is more skilled than the arbitrary zerg player. This implies the race is stronger than zerg (and to a lesser extent, my brethren from Aiur).
|
Guys, the matchmaking is designed to put players against others so that they will win around 50% of the time. If one race took 10 APM to play and the other took 300, there would still be a 50% win/loss ratio.
|
Win rates don't mean anything about balance. In a matchup, you can have a very overpowered early middle game against a race very overpowered late game. You can obtain a 50% win ratio in the end, but with disgustingly imbalanced and not interesting games.
|
On October 08 2010 11:31 nickjpgeorge wrote:The first graph here is 100x more telling than any of Blizz's stats: http://www.sc2ranks.com/stats/race/us/1/The very clear correlation between increase in Diamond points and increase in Terran percentage means one thing: According to the match-making system, the arbitrary Terran player is more skilled than the arbitrary zerg player. This implies the race is stronger than zerg (and to a lesser extent, my brethren from Aiur).
Well I dont know, when looking at those graphs it seems the "overall" race chosen in all leagues (besides bronze) is Protoss. What does that mean? Sure there are no high-level diamonds (2200+) using Protoss or Zerg, but how can you infer anything from that?
Kudos on the link, I like seeing things in pretty graphs and colours!
EDIT: Oh when I look at http://www.sc2ranks.com/ranks/us it shows the sample size, there are only 3 players above 2200 and 8 players above 2100. So these "top tier stats" aren't that really effective of convincing us of anything
|
I wonder whether Blizzard analyzes correlation between build order and win % in determining what units to target with their nerf bat. For instance if they saw that in Diamond ladder games that in TvZ if the T gets early reapers that they win 75% of the time that might account for them targeting that particular unit.
|
On October 08 2010 07:48 Crabman123 wrote:
• We're increasing roach range. This will allow roaches to be more effective in large groups, giving the zerg more options in the mid to end game.
• Fungal Growth will now prevent Blink, which will give zerg a way to stop endlessly Blinking stalkers which can be very challenging to deal with in large numbers.
• The Barracks are going to require a Supply Depot, which will impact a lot of early terran reaper pushes.
• The reaper speed upgrade will require the Factory, which is meant to weaken a lot of the early terran reaper attacks that dominate so many matches, especially in team games.
• We're making a number of increases to the health of zerg buildings, which will make the very vulnerable zerg technology structures more resistant to raids. We don't expect these hit point changes to have a super significant impact on the game, but the current numbers felt way too low.
Gosh blizz thank you for letting me go 14 hatch every game against T.
Sorry, but wtf is this? This just has been randomly posted in this thread, and noone has mentioned wtf it is or where it's from.
I just thought this guy was trolling. What is this shit?
|
On October 08 2010 12:29 Subversion wrote:
Sorry, but wtf is this? This just has been randomly posted in this thread, and noone has mentioned wtf it is or where it's from.
I just thought this guy was trolling. What is this shit?
Developer's Corner: 1v1 Game Balance - StarCraft II
http://us.battle.net/sc2/en/blog/882511
|
On October 08 2010 12:29 Subversion wrote:Show nested quote +On October 08 2010 07:48 Crabman123 wrote:
• We're increasing roach range. This will allow roaches to be more effective in large groups, giving the zerg more options in the mid to end game.
• Fungal Growth will now prevent Blink, which will give zerg a way to stop endlessly Blinking stalkers which can be very challenging to deal with in large numbers.
• The Barracks are going to require a Supply Depot, which will impact a lot of early terran reaper pushes.
• The reaper speed upgrade will require the Factory, which is meant to weaken a lot of the early terran reaper attacks that dominate so many matches, especially in team games.
• We're making a number of increases to the health of zerg buildings, which will make the very vulnerable zerg technology structures more resistant to raids. We don't expect these hit point changes to have a super significant impact on the game, but the current numbers felt way too low.
Gosh blizz thank you for letting me go 14 hatch every game against T. Sorry, but wtf is this? This just has been randomly posted in this thread, and noone has mentioned wtf it is or where it's from. I just thought this guy was trolling. What is this shit?
It's posted at the bottom of Browder's blog post, linked from the front page on bnet.
Or, at least it is on NA site: link
Edit: Beaten to it.
|
On October 08 2010 12:20 Strutswell wrote:Show nested quote +On October 08 2010 11:31 nickjpgeorge wrote:The first graph here is 100x more telling than any of Blizz's stats: http://www.sc2ranks.com/stats/race/us/1/The very clear correlation between increase in Diamond points and increase in Terran percentage means one thing: According to the match-making system, the arbitrary Terran player is more skilled than the arbitrary zerg player. This implies the race is stronger than zerg (and to a lesser extent, my brethren from Aiur). Well I dont know, when looking at those graphs it seems the "overall" race chosen in all leagues (besides bronze) is Protoss. What does that mean? Sure there are no high-level diamonds (2200+) using Protoss or Zerg, but how can you infer anything from that? Kudos on the link, I like seeing things in pretty graphs and colours! EDIT: Oh when I look at http://www.sc2ranks.com/ranks/us it shows the sample size, there are only 3 players above 2200 and 8 players above 2100. So these "top tier stats" aren't that really effective of convincing us of anything
True. But I'm more interested in the 1600-1900 range, where there are a statistically significant number of players. Protoss flops around a lot, so I don't think there's any inferrable trend, but Terran clearly goes up and zerg clearly goes down.
|
"The balance changes in our next patch will primarily focus on improving the zerg." I just want to weep with happiness when I read that.
Also, why is there no thread on this? Or is there? O_o
|
On October 08 2010 11:24 hdkhang wrote:Show nested quote +On October 08 2010 10:32 Techno wrote:On October 08 2010 10:02 Hypatio wrote:
I'm zerg and my win ratio is about 53% at about 1k points. I win maybe 75% ZvZ's, win about 50% of ZvP and win maybe 25% of my ZvT's. You should work on your ZvT On October 08 2010 10:30 Adeny wrote:On October 08 2010 10:22 Techno wrote:
I dont see how people can say these stats are meaningless without also saying other stats like "4 hydras beat 3 marauders" is meaningless. These stats mean something. They mean that Zerg players are beating Terran players. Something I think most of us are learning to accept. Just a vocal minority left....
They mention in the article that they are considering ways of nerfing Protoss to alter Protoss vs Terran late game. Although I do not cry of imbalance, I am open to changes in Protoss vs Terran that favor T when the game reaches what I call "macro stages". They are meaningless because you don't know wtf they portray, it could literally be anything. 4 hydras beating 3 marauders is also completely meaningless without additional information and I don't think you could find anyone to disagree with that. They portray exactly what they say they portray. Race vs Race win/loss ratios at all the different leagues. What kind of statistic DOES mean something to you? To me this means that the game is pretty damn balanced, of course it isnt perfect, but its pretty damn balanced. How can anyone disagree? By disagree you would say that the game is very unbalanced, and to me that means that it is possible for a player to win 100% of his/her games by doing a particular thing which has no response. I'm sure players will say things like "MMM has no response", but that's getting really old and I will not waste my cyber breath on heathens such as that. The stats only show that the Matchmaking system is working. The only stat that blizzard collects is the match outcome i.e. wins and losses. Based on your wins and losses vs other players wins and losses, you get somewhat ranked, this is their hidden rating.
Did you not even read the OP? The stats are adjusted for skill, insofar as the MMR system allows this to be done.
It is possible to do this, with some degree of accuracy.
|
On October 08 2010 10:21 eloist wrote:
I don't think people as proficient at writing a match making system that works this well wouldn't be able to point out to their designer colleagues which factor are affecting the statistics in what why and how to filter it for reporting.
Trust me, the "good" matchmaking system has more to do with good programing, which we already know they master, than with advanced statistical evaluation.
Their method is basically like this: They rate all players based on a formula similar to the ELO system. The match-making system then force BattleNet to pair players based on their rating relative to other players. The larger the difference in rating between two players, the lesser likelihood of the two being paired. Voala!
The WC3 match-making system was actually pretty bad, as it didn't track how large of an population was online at one time. This resulted in some higher rated players being "forbidden" to play as their was no-one around matching their rating. They seem to have fixed this for SC2.
On October 08 2010 10:39 RoboBob wrote:
This stuff is pretty interesting.
Did blizzard ever say why their "hidden rating" is hidden?
Well, probably because it would undermine the legitimity of their visible division-ranking systems. This second "superficial" ranking system is in place as they believe it is better at motivating people to keep playing.
On October 08 2010 11:23 tGhOeOoDry wrote:
-The key problem with the statistics has already been mentioned: the hidden player skill value is determined by your wins and losses. Thus, if a racial imbalance exists, Zerg players will lose more often to terrans and will appear to be less skillful. That means that there is correlation between the independent and dependent variables in the regression, which makes the conclusions of the regression suspect.
Well, you can still come too a conclusion on which race is inferior if you assume the same distribution of skill or talent across all three races .
On the other hand you can assume different distribution of skill or talent across the races. In this case you can still use the same method as long as future beginners of each race will follow the same race-specific skill distribution as the current ones.
The issue isn't really the data, my concern is how Blizzard chose to present it.
|
aaah man im so glad you guys all see it. win% stats are doctored by the automatch system and are therefore meaningless. We are too smart for you blizzard, you cant pull that wool over our eyes :D I would suggest using racial distribution as a gauge to balance instead.
|
On October 08 2010 13:45 trueg0x wrote:
... I would suggest using racial distribution as a gauge to balance instead.
lol 
|
On October 08 2010 07:32 Pyrrhuloxia wrote:
They show what I suspected, that PvZ is the most imbalanced matchup right now.
exactly what I expect from a 5k poster.
Wait what?
You know this has nothing to do with imbalance right, its matchmaking, showing nothing about balance.
|
On October 08 2010 07:31 Fa1nT wrote:
+/- 3% means matchmaking is doing its job and that these statistics are pointless?
Explain protoss 60% win vs T/Z everywhere but high diamond.
what excuse we gonna hear?
|
On October 08 2010 13:45 trueg0x wrote:
aaah man im so glad you guys all see it. win% stats are doctored by the automatch system and are therefore meaningless. We are too smart for you blizzard, you cant pull that wool over our eyes :D I would suggest using racial distribution as a gauge to balance instead.
On October 08 2010 13:50 Apexplayer wrote:Show nested quote +On October 08 2010 07:32 Pyrrhuloxia wrote:
They show what I suspected, that PvZ is the most imbalanced matchup right now. exactly what I expect from a 5k poster. Wait what? You know this has nothing to do with imbalance right, its matchmaking, showing nothing about balance.
I seriously want to strange all these kids who think they are smarter than everyone else posting this garbage not reading the fucking post carefully, or any of the responses in the thread. FDASJKLFSJKL. Here, kzn explains it.
On October 08 2010 13:11 kzn wrote:Show nested quote +On October 08 2010 11:24 hdkhang wrote:On October 08 2010 10:32 Techno wrote:On October 08 2010 10:02 Hypatio wrote:
I'm zerg and my win ratio is about 53% at about 1k points. I win maybe 75% ZvZ's, win about 50% of ZvP and win maybe 25% of my ZvT's. You should work on your ZvT On October 08 2010 10:30 Adeny wrote:On October 08 2010 10:22 Techno wrote:
I dont see how people can say these stats are meaningless without also saying other stats like "4 hydras beat 3 marauders" is meaningless. These stats mean something. They mean that Zerg players are beating Terran players. Something I think most of us are learning to accept. Just a vocal minority left....
They mention in the article that they are considering ways of nerfing Protoss to alter Protoss vs Terran late game. Although I do not cry of imbalance, I am open to changes in Protoss vs Terran that favor T when the game reaches what I call "macro stages". They are meaningless because you don't know wtf they portray, it could literally be anything. 4 hydras beating 3 marauders is also completely meaningless without additional information and I don't think you could find anyone to disagree with that. They portray exactly what they say they portray. Race vs Race win/loss ratios at all the different leagues. What kind of statistic DOES mean something to you? To me this means that the game is pretty damn balanced, of course it isnt perfect, but its pretty damn balanced. How can anyone disagree? By disagree you would say that the game is very unbalanced, and to me that means that it is possible for a player to win 100% of his/her games by doing a particular thing which has no response. I'm sure players will say things like "MMM has no response", but that's getting really old and I will not waste my cyber breath on heathens such as that. The stats only show that the Matchmaking system is working. The only stat that blizzard collects is the match outcome i.e. wins and losses. Based on your wins and losses vs other players wins and losses, you get somewhat ranked, this is their hidden rating. Did you not even read the OP? The stats are adjusted for skill, insofar as the MMR system allows this to be done. It is possible to do this, with some degree of accuracy.
|
It's sort of obnoxious that people are discussing this and acting like they know what's going on and how to interpret this data, when we don't know anything about how it was calculated or the details of how the matchmaking system really works. All we can do for now is accept Blizzard's interpretations. It drives me crazy that so many internet experts think that they know what's going on and that Blizzard doesn't. It's also frustrating that the minute they release what could be helpful data, people jump down their throats.
|
I wonder if Terran was 52% vs x people would still be questioning the stats. Actually, I don't think I have to wonder: regardless of the truth (as proven by Blizzard, as proven elsewhere) the whiners will stick with the false narrative. Nothing will convince these folks otherwise -- repeating a lie long enough does that. But more and more we see that the game is very balanced and if any one race is OP, it is most certainly Protoss.
|
Hopefully no one has mentioned this. I read most of the posts but started skipping some once everyone started yelling at each other. One of the only accurate measurements of skill that would exclude race balance/imbalance would be to measure skill based on the percentage of wins of mirror matchups. So say if I am a 1000 platinum zerg who wins 60% of my mirro matchups. And then I beat a terran who only wins about 40% of their mirror matchups, then my hidden skill ranking (that blizzard says they are using) would not be as high as if I beat a terran who won 60% of their mirror.
I'm really tired so hopefully what I said isn't too confusing. But there really seems like very few ways if any that Blizzard can claim that those stats are adjusted for skill. There is just too many people with way too many variables.
The only other thing, I think they could use would be APM, but I don't think many people would agree that is a great indicator for skill.
Oh well. I am glad they are at least reading the forums and understand that people aren't happy with the way things are, even if the stats they give us are complete garbage.
|
On October 08 2010 13:51 klauz619 wrote:Show nested quote +On October 08 2010 07:31 Fa1nT wrote:
+/- 3% means matchmaking is doing its job and that these statistics are pointless?
Explain protoss 60% win vs T/Z everywhere but high diamond. what excuse we gonna hear?
proof?
If this was a known fact im fairly certain there would be some outrage/discussion. As is theres nothing that says you didnt pull those numbers out of your ass.
Basically you show us proof that toss wins 60% of games out of high diamond, then maybe people will answer
|
On October 08 2010 14:31 Chen wrote:Show nested quote +On October 08 2010 13:51 klauz619 wrote:On October 08 2010 07:31 Fa1nT wrote:
+/- 3% means matchmaking is doing its job and that these statistics are pointless?
Explain protoss 60% win vs T/Z everywhere but high diamond. what excuse we gonna hear? proof? If this was a known fact im fairly certain there would be some outrage/discussion. As is theres nothing that says you didnt pull those numbers out of your ass. Basically you show us proof that toss wins 60% of games out of high diamond, then maybe people will answer
Wow did you even read the SC2 statistics dustin browder put up along with some of the patch notes?
|
One of the most interesting/important things I learned in economics is to look at what people do, not what people say. Obviously even Blizzard knows these statistics mean nothing or they would notsay they are going to make so many changes to zerg. (unless they truly don't care about balance, which I guess is possible as well)
|
On October 08 2010 13:57 AcrossFiveJulys wrote:Show nested quote +On October 08 2010 13:45 trueg0x wrote:
aaah man im so glad you guys all see it. win% stats are doctored by the automatch system and are therefore meaningless. We are too smart for you blizzard, you cant pull that wool over our eyes :D I would suggest using racial distribution as a gauge to balance instead. Show nested quote +On October 08 2010 13:50 Apexplayer wrote:On October 08 2010 07:32 Pyrrhuloxia wrote:
They show what I suspected, that PvZ is the most imbalanced matchup right now. exactly what I expect from a 5k poster. Wait what? You know this has nothing to do with imbalance right, its matchmaking, showing nothing about balance. I seriously want to strange all these kids who think they are smarter than everyone else posting this garbage not reading the fucking post carefully, or any of the responses in the thread. FDASJKLFSJKL. Here, kzn explains it. Show nested quote +On October 08 2010 13:57 AcrossFiveJulys wrote:On October 08 2010 13:45 trueg0x wrote:
aaah man im so glad you guys all see it. win% stats are doctored by the automatch system and are therefore meaningless. We are too smart for you blizzard, you cant pull that wool over our eyes :D I would suggest using racial distribution as a gauge to balance instead. On October 08 2010 13:50 Apexplayer wrote:On October 08 2010 07:32 Pyrrhuloxia wrote:
They show what I suspected, that PvZ is the most imbalanced matchup right now. exactly what I expect from a 5k poster. Wait what? You know this has nothing to do with imbalance right, its matchmaking, showing nothing about balance. I seriously want to strange all these kids who think they are smarter than everyone else posting this garbage not reading the fucking post carefully, or any of the responses in the thread. FDASJKLFSJKL. Here, kzn explains it. On October 08 2010 13:11 kzn wrote:On October 08 2010 11:24 hdkhang wrote:On October 08 2010 10:32 Techno wrote:On October 08 2010 10:02 Hypatio wrote:
I'm zerg and my win ratio is about 53% at about 1k points. I win maybe 75% ZvZ's, win about 50% of ZvP and win maybe 25% of my ZvT's. You should work on your ZvT On October 08 2010 10:30 Adeny wrote:On October 08 2010 10:22 Techno wrote:
I dont see how people can say these stats are meaningless without also saying other stats like "4 hydras beat 3 marauders" is meaningless. These stats mean something. They mean that Zerg players are beating Terran players. Something I think most of us are learning to accept. Just a vocal minority left....
They mention in the article that they are considering ways of nerfing Protoss to alter Protoss vs Terran late game. Although I do not cry of imbalance, I am open to changes in Protoss vs Terran that favor T when the game reaches what I call "macro stages". They are meaningless because you don't know wtf they portray, it could literally be anything. 4 hydras beating 3 marauders is also completely meaningless without additional information and I don't think you could find anyone to disagree with that. They portray exactly what they say they portray. Race vs Race win/loss ratios at all the different leagues. What kind of statistic DOES mean something to you? To me this means that the game is pretty damn balanced, of course it isnt perfect, but its pretty damn balanced. How can anyone disagree? By disagree you would say that the game is very unbalanced, and to me that means that it is possible for a player to win 100% of his/her games by doing a particular thing which has no response. I'm sure players will say things like "MMM has no response", but that's getting really old and I will not waste my cyber breath on heathens such as that. The stats only show that the Matchmaking system is working. The only stat that blizzard collects is the match outcome i.e. wins and losses. Based on your wins and losses vs other players wins and losses, you get somewhat ranked, this is their hidden rating. Did you not even read the OP? The stats are adjusted for skill, insofar as the MMR system allows this to be done. It is possible to do this, with some degree of accuracy. If anyone familiar with regression could explain to me how you can perform regression when you have no idea what form any of the variables take even. As far as im concerned you just have a set of points that are a function of Race and True skill where true skill is not expressable as noone even knows what makes it up let alone how to quantify it. Then I think you have to know how the variables relate atleast loosely. I think it could be quite easy to make a bad regression to give you numbers you want to see. I think this is an interesting way to get a better understand of regressional analysis so I pose my challenge more as a way to hear a response from someone with more experience. Show nested quote +On October 08 2010 13:11 kzn wrote:On October 08 2010 11:24 hdkhang wrote:On October 08 2010 10:32 Techno wrote:On October 08 2010 10:02 Hypatio wrote:
I'm zerg and my win ratio is about 53% at about 1k points. I win maybe 75% ZvZ's, win about 50% of ZvP and win maybe 25% of my ZvT's. You should work on your ZvT On October 08 2010 10:30 Adeny wrote:On October 08 2010 10:22 Techno wrote:
I dont see how people can say these stats are meaningless without also saying other stats like "4 hydras beat 3 marauders" is meaningless. These stats mean something. They mean that Zerg players are beating Terran players. Something I think most of us are learning to accept. Just a vocal minority left....
They mention in the article that they are considering ways of nerfing Protoss to alter Protoss vs Terran late game. Although I do not cry of imbalance, I am open to changes in Protoss vs Terran that favor T when the game reaches what I call "macro stages". They are meaningless because you don't know wtf they portray, it could literally be anything. 4 hydras beating 3 marauders is also completely meaningless without additional information and I don't think you could find anyone to disagree with that. They portray exactly what they say they portray. Race vs Race win/loss ratios at all the different leagues. What kind of statistic DOES mean something to you? To me this means that the game is pretty damn balanced, of course it isnt perfect, but its pretty damn balanced. How can anyone disagree? By disagree you would say that the game is very unbalanced, and to me that means that it is possible for a player to win 100% of his/her games by doing a particular thing which has no response. I'm sure players will say things like "MMM has no response", but that's getting really old and I will not waste my cyber breath on heathens such as that. The stats only show that the Matchmaking system is working. The only stat that blizzard collects is the match outcome i.e. wins and losses. Based on your wins and losses vs other players wins and losses, you get somewhat ranked, this is their hidden rating. Did you not even read the OP? The stats are adjusted for skill, insofar as the MMR system allows this to be done. It is possible to do this, with some degree of accuracy.
If anyone familiar with regression could explain to me how you can perform regression when you have no idea what form any of the variables take even. As far as im concerned you just have a set of points that are a function of Race and True skill where true skill is not expressable as noone even knows what makes it up let alone how to quantify it. Then I think you have to know how the variables relate atleast loosely. I think it could be quite easy to make a bad regression to give you numbers you want to see.
I think this is an interesting way to get a better understand of regressional analysis so I pose my challenge more as a way to hear a response from someone with more experience.
Edit: Someone above mentioned counting mirror matches as a way of finding true skill but that fails to account for some people being better at specific matchups. Someone could be terrible at the mirror and then be the best in the world at the other two matchups.
|
On October 08 2010 07:44 Liquid`Tyler wrote:Show nested quote +On October 08 2010 07:41 nemahsys wrote:
I want to know how they are "accounting for player skill" when coming up with these numbers. Not really sure how you go about doing that. That's essentially the question that is on the table already, the whole point of this thread.
I'd say it's simply not possible - believing it, is an illusion
just for the sake of argument, let's say protoss has a MASSIVE advantage over zerg; then a zerg-player must be of much higher skill to beat a protoss-opponent;
this basicly means, that in this case a 1k point zerg may have a 50/50 win-ratio vs a 1k point protoss, but will be much higher skilled; the system is designed to get you a 50/50 ratio, which means I seriously doubt that blizzard can "account for skill" in these statistics;
because it would just mean, forgetting about the system overall; if you have two players of different races with equal amount of points and equal amount of win-% there is NO WAY to tell, if not one of these two players is much higher skilled but ended up "lower" because of his racial disadvantage
|
On October 08 2010 07:31 Fa1nT wrote:
+/- 3% means matchmaking is doing its job and that these statistics are pointless?
well said
|
The only way to figure out balance is through many many many highest lvl games played in a competetive environment. Obviously there havent been enough tournaments yet so far, and the problem will always be that the samplesize is small by the definition of which data counts and which does not.
|
attention: "adjusted for skill level" means the stats posted compare only evenly matched opponents, i.e., opponents with same hidden skill level. this means the stats ARE NOT a reflection of the ladder system, which pits players against lower and higher skilled players to drive a player to 50% win ratio.
in sum, PLEASE STOP saying that the stats are only a reflection of the ladder system working.
|
Here's an easy way to determine player skill:
Count how many mirror matches the player wins = how skilled the player is =)
If the good zerg players (who win 60-70% of all mirror matches in diamond) then lose consistently to protoss, it would suggest PvZ is imbalanced in favour of P etc.
Note, I said suggest (it might also mean that people who are good at ZvZ are bad at ZvP!), but that seems less likely over a large enough sample.
|
On October 08 2010 07:51 TitleRug wrote:Show nested quote +On October 08 2010 07:41 nemahsys wrote:
I want to know how they are "accounting for player skill" when coming up with these numbers. Not really sure how you go about doing that. The people at Blizzard are probably smart enough to create an algorithm to do this. I hope so at least .
I takes much a priori information to make such an algorithm. One I could think of would be : for every race, the skill distribution is the same among the players. It would be very great though if they told us how they did it precisely, because as such the data is completely useless.
I can too tell number 1223 552 21.132 see ?
|
On October 08 2010 17:36 Svetz wrote:
Here's an easy way to determine player skill:
Count how many mirror matches the player wins = how skilled the player is =)
wtf why? I could be a good player but just suck at my mirror - or eg. I even start to cheese in PvP once in a while because it's SO annoying to get like the 5th mirror-matchup in a row, don't care anymore if I lose, just want to end it quickly
|
On October 08 2010 17:39 sleepingdog wrote:Show nested quote +On October 08 2010 17:36 Svetz wrote:
Here's an easy way to determine player skill:
Count how many mirror matches the player wins = how skilled the player is =)
wtf why? I could be a good player but just suck at my mirror - or eg. I even start to cheese in PvP once in a while because it's SO annoying to get like the 5th mirror-matchup in a row, don't care anymore if I lose, just want to end it quickly
Read the second half of the post, hence why you need a large sample =)
|
dezi
Germany1536 Posts
On October 08 2010 07:32 Pyrrhuloxia wrote:
They show what I suspected, that PvZ is the most imbalanced matchup right now.
49.6% (PvT) seems more balanced to me then 52.8% (PvZ).
|
On October 08 2010 13:51 klauz619 wrote:Show nested quote +On October 08 2010 07:31 Fa1nT wrote:
+/- 3% means matchmaking is doing its job and that these statistics are pointless?
Explain protoss 60% win vs T/Z everywhere but high diamond. what excuse we gonna hear?
Rushes. Not needing anything but a probe to build stuff makes it very easy to outright win from the start. And probably 4 Gate. Every time I play against protoss I am super paranoid and I think 20% of the games were proxy 2gates, in base 2gates, cannon rushes, cannons@ramp and it takes a bit of game understanding to break these if they managed to start them and they know where to place stuff.
|
On October 08 2010 07:41 nemahsys wrote:
I want to know how they are "accounting for player skill" when coming up with these numbers. Not really sure how you go about doing that.
As far as I'm aware, it's completely impossible. Blizzard's only measure of skill is your MMR, which is also what ensures that you get a 50% win ratio. There's no way they could separate the two.
|
now when protoss/zerg leads in statistics everybody says its fine and its match making system fault but when Terran led marginally there were countless of whine posts and people claiming imba ...
|
I am completely unsure of how their MMR systems takes into account race-based imbalances. Race based imbalances are incredibly difficult to tease out from statistical analysis of a a system where match-making is based on skill rating, which is itself based on win percentage. I am actually very mathematically curious about this.
|
As I understand it, most people, especially Zerg players, aren't actually complaining that their race is inferior but that it is more difficult to play at certain skill levels. Honestly I don't understand why people make such a big fuss about that. It was the same in Broodwar and it also has a lot to do with currently popular strategies and discoveries of high level players(e.g. magic box).
While it is a valiant effort from Blizzard to try and balance the game at every level, I think it is almost impossible. I understand this has to suck for Bronze/Silver/Gold/Plat/Low Diamond players, but you have to understand that this is a competitive, e-sports oriented game and thus has to be balanced at the very highest level of play. The level I'm talking about most likely hasn't even been touched yet by any SC2 "progamer".
What's my point? If you're in Bronze, Silver, Gold, Platinum, or low level Diamond, don't even THINK about balance. You have so many aspects of your game to improve that balancing is the least of your problems. Even if you're mid to high level Diamond without reaching a competitive level of play, balance shouldn't concern you much. While your understanding of the game might be good enough to reasonably think about whether something might be a touch overpowered at the highest level, you're probably still miles away from that level.
Imho Blizzard did a damn good job already at balancing this game across all the various levels of play and that's exactly what these recently released numbers show. Now they should focus on high level balancing, and not stuff that low level players feel to be imbalanced because it's simply a bit easier to execute than what their own race can do at their respective level.
Edit: To all those saying that these statistics don't show the actual skill etc. Imho skill is what you are able to do with your race. That's your current skill level. Thus the matchmaking system matches you with players of the same skill, at least by my definition. If you feel like another race would be more rewarding to play with your abilities, feel free to switch and see if it's true. Sometimes what makes a race "easier" to play doesn't have anything to do with the actual mechanics. It's just that it suits your personal style better.
|
They never said they use MMR for anything other than matchmaking. The statistics they posted do not say they come from anything related to MMR. MMR is a statistic they make from 1 player's game data and use it to go towards a 50% win ratio. These on the other hand can come after filtering for any number of anomalies (they do have access to extreme amounts of raw data about all the games played on the ladder).
While overall the zerg buffs are quite nice they can simply remove reaper from the game. Not seen anyone starting to make reapers mid-late game yet, stimmed marauders are much more reliable and don't die instantly.
Barracks requiring supply means a bit longer reliance on just SCVs to fight off rushes, just solved by being even slightly more paranoid now and scouting your base and surroundings way earlier, it's a bit difficult to stop a proxy smth if the buildings have gone over 25% with just SCVs, especially protoss that blocks pylon access to just 1 unit.
They do need to do something about Protoss rushes though if they ever want to balance the games < diamond - though at least to me it didn't seem that bad, it's actually funny to watch the rushers curse their way out of a game with anti-terran insults after they massively failed; it only takes a bit of paranoia, a bit of luck and good placement of troops/buildings.
|
Also, as many players that initialy played zerg switched race, it is more probable that those who sticked with zerg will be the ones that had positive win ratio vs t (before patch). So the result will be slightly favored towards z.
|
Surely it is better to go by aggregate tourney results rather than these dubious stats produced from a system that tries to auto-balance players. I have seen many absolutely atrocious (skill-wise) terrans getting reasonably high diamond (terrible, terrible macro compared to toss and zerg counterparts of the same matchmaking). Stats like Terran winning 12 consecutive Zotac cups is what should be looked at by blizzard. Balancing based on the highest level of play will hardly impact the OP's stats imo and seems like the better way to proceed.
|
What is with people finding crazy ideas as to what these statistics mean. I'm really supsrised no one went for the obvious explanation, it just kind of doesn't make sense untill you think about it abit. This is the way I'm reading the statistics, my opinion is actually differant as I think Terran are easier to play. Also I'm assuming the sample size is big and it's not biased in anyway because that changes everything.
I'm only going to do TvP because it'll just get too complicated for a simple concept even if it doesn't feel right.
The simple answer is Protoss is easier to play at equal skill levels when the skill level is low. But when the skill level evens out to a certain degree at higher levels apparently, Terran start becoming over powered over Protoss. Now think about how close the stats are in PvT at dimond, but how many Terran win tornmanets? Well it makes sense if Terran are worse for less skillful players to play and more of an advantage for more skillfull players to play. This advantage should be widening the higher up you go making Terran easier for very top level play. Also with the statistics if you go into the top percentile it should show more Terran in the top 100 or what ever and if you look at the ranks it does.
Look at it this way too. Terran in my opinion have an advantage early game against Protoss and Protoss have the advantage late game.Lower level players don't understand the importance of early pressure so less of them will do it or if they do not as effective, which could be one of the factors that make Protoss easier to play for lower levels.
|
Great!!! At last blizzard balanced the game perfectly
|
Jesus, when will people stop caring about win ratio, which means only that the ladder system is working as intended, and nothing more...
If you want to check for imbalances, check the repartition of the races in terms of points, not win ratio -_- ...
|
Points are irrelevant in terms of skill. The top end players will always get more points than everyone else since there aren't enough top end players to be matched with at any time. So the system will get games against lower rated opponents quite a bit more than a mid level player that will get somewhat even opponents.
Win ratio if you remove games that are clearly in your favor or clearly favored to lose actually has some relevance. The matchmaking system will give you some of those opponents because it doesn't have an infinite pool to choose from but that doesn't mean Blizzard has to use those results, they're not limited to only what you see on your screen, remember that.
|
Surely it is better to go by aggregate tourney results rather than these dubious stats produced from a system that tries to auto-balance players. I have seen many absolutely atrocious (skill-wise) terrans getting reasonably high diamond (terrible, terrible macro compared to toss and zerg counterparts of the same matchmaking). Stats like Terran winning 12 consecutive Zotac cups is what should be looked at by blizzard. Balancing based on the highest level of play will hardly impact the OP's stats imo and seems like the better way to proceed.
That's even worse IMO because then you start to balance around the handful of players that are capable of winning tournaments. And if you do that, you might as well say Zerg is fine because a Zerg player won the biggest Sc2 tournament to-date.
|
Jesus, when will people stop caring about win ratio, which means only that the ladder system is working as intended, and nothing more...
If you want to check for imbalances, check the repartition of the races in terms of points, not win ratio -_- ...
Because this is the first time that Blizzard expressly says they accounted for the system pushing people to 50% when producing the stats.......
|
On October 08 2010 19:51 starckr wrote:Show nested quote +Jesus, when will people stop caring about win ratio, which means only that the ladder system is working as intended, and nothing more...
If you want to check for imbalances, check the repartition of the races in terms of points, not win ratio -_- ... Because this is the first time that Blizzard expressly says they accounted for the system pushing people to 50% when producing the stats.......
Blizzard saying they did it doesn't mean much, when there's no conceivable way that they could actually do it.
I love the way people just say "Oh yeah, Blizzard pulled out numbers from their vast raw data from games played!", like there's some mystical data hidden in each game played, which can tell you how skilled a player is, while ignoring who won the game.
Blizzard's only way to determine your skill is with MMR, and MMR is also what gives you a 50% win ratio. How then does Blizzard possibly compensate for the 50% win ratios?
Vast amounts of RAW UNTAPPED MAGIC -- er -- DATA!!
|
Blizzard saying they did it doesn't mean much, when there's no conceivable way that they could actually do it.
I love the way people just say "Oh yeah, Blizzard pulled out numbers from their vast raw data from games played!", like there's some mystical data hidden in each game played, which can tell you how skilled a player is, while ignoring who won the game.
Blizzard's only way to determine your skill is with MMR, and MMR is also what gives you a 50% win ratio. How then does Blizzard possibly compensate for the 50% win ratios?
Vast amounts of RAW UNTAPPED MAGIC -- er -- DATA!!
Huh? What drives people to 50% is that the system pairs you against people you are likely to lose against or win against to try and drive you to 50%. Now, along the way, you play people that are both higher in skill, lower in skill, and the same skill. Blizzard eliminates the higher skill / lower skill games to arrive at games involving equally skilled opponents. That stat has at least more value than the overall stats, which unquestionably drive you to 50% because it includes games vs. lower and higher MMR.
The question of how much value is in it is debatable, but saying it only represents the system that drives people to 50% is clearly false. That is a function of the system pairing you vs. stronger and weaker opponents.
|
On October 08 2010 20:41 starckr wrote:Show nested quote +Blizzard saying they did it doesn't mean much, when there's no conceivable way that they could actually do it.
I love the way people just say "Oh yeah, Blizzard pulled out numbers from their vast raw data from games played!", like there's some mystical data hidden in each game played, which can tell you how skilled a player is, while ignoring who won the game.
Blizzard's only way to determine your skill is with MMR, and MMR is also what gives you a 50% win ratio. How then does Blizzard possibly compensate for the 50% win ratios?
Vast amounts of RAW UNTAPPED MAGIC -- er -- DATA!! Huh? What drives people to 50% is that the system pairs you against people you are likely to lose against or win against to try and drive you to 50%. Now, along the way, you play people that are both higher in skill, lower in skill, and the same skill. Blizzard eliminates the higher skill / lower skill games to arrive at games involving equally skilled opponents. That stat has at least more value than the overall stats, which unquestionably drive you to 50% because it includes games vs. lower and higher MMR. The question of how much value is in it is debatable, but saying it only represents the system that drives people to 50% is clearly false. That is a function of the system pairing you vs. stronger and weaker opponents.
No. If you take only the players of the same 'skill' (not favored), and say this is the win ratio when you remove the 50% phenomenon, you are wrong. It is a result of the system of 50% win ratio that some players are around your hidden mmr, so all the results are biased. Statistically, if the system works well and you make only even matches, you should have 50%.
I don't how they can account for it, or if it is even possible (I doubt it), but if they do like you say it's definitely wrong. But then we will never know since they won't release their way of calculating it.
|
There are AFAIK at least two possible ways to do this, mentionned in this thread so far :
- Assume that the skill distribution among every player is the same for the three races (you can then infer the skill of an individual player knowing its MMR, its race, and the skill distribution).
- Consider mirror match-ups to compute a skill rating, independant of the race ; For inter racial comparison, this works, though, only if the skill distribution is the same for the three races now that I think about it.
I guess this leaves us with one possibility, quite dubious in my opinion.
|
What would be nice to see is some statistics from online and ofline tourneys MU.
There was a discussion at some point regarding this but it would be nice to see the stats remade to date.This would be made with public data so that would be a starting point of discussion witch we dont have here.
Even in online armature tourneys the field is mostly made of top diamonds.Maybe from this only the last rounds can be taken in consideration where the level of play increases and the difference in skill disappears.
As was said before looking at the games at the higher level of play is the only way to determine balance because in the lower levels( battlenet ladder) the data is irelevant due to different factors(the match making system).
|
I don't think you can really infer anything about balance from ladder. The matchmaking system (or my understanding of it) doesn't take race into consideration when matching opponents. It just looks at relative score and tries to match you against someone who is close to your score, then adjusts your score after each match based on the outcome.
So a relatively low win % for any race could just be that the matchmaking system for those people is more accurate in matching opponents of similar skill levels, or that the matchmaking system for other races might just be placing those players into easy matches that they are likely to win.
To understand game balance it's far more useful to look at tournaments and individual games/tactics. For the game to be balanced, then at all points the two players must be able to combat each others strategies and tactics in such a way that the player who is better able to control his units, timings, and scouting will win. If there is any point in a game where race x can do y and win every game if he controls his units well, then that is imbalance and can't really be shown by a ladder system that will match opponents based on who they are likely to win/lose to.
|
On October 08 2010 18:04 Cashout wrote:
now when protoss/zerg leads in statistics everybody says its fine and its match making system fault but when Terran led marginally there were countless of whine posts and people claiming imba ...
Yeah thats quite ridiculous to say the least.. Even when the "almighty" FruitSeller explains that tvz is fine for him since 1.1, they still cry at imbaness.
Honestly its really boring, every RTS I played I always faced crazy talk on the forums and endless streams of whine everywhere.
If the MMR whatever force the 50% win, because like in chess lets say that terran (white) have a +Xpoints MMR due to racial imbalance, why do we see stats like 55-45 for zerg or 60-40 for toss in some leagues ? Could someone explain that ?
|
On October 08 2010 19:50 starckr wrote:Show nested quote +Surely it is better to go by aggregate tourney results rather than these dubious stats produced from a system that tries to auto-balance players. I have seen many absolutely atrocious (skill-wise) terrans getting reasonably high diamond (terrible, terrible macro compared to toss and zerg counterparts of the same matchmaking). Stats like Terran winning 12 consecutive Zotac cups is what should be looked at by blizzard. Balancing based on the highest level of play will hardly impact the OP's stats imo and seems like the better way to proceed. That's even worse IMO because then you start to balance around the handful of players that are capable of winning tournaments. And if you do that, you might as well say Zerg is fine because a Zerg player won the biggest Sc2 tournament to-date.
That's just wrong, is it just coincidence that more of the players capable of winning tourneys picked terran? If blizz doesn't balance based upon the pro's then SC2 will fail as an esport. Overall high-end tourney results do tend to show racial imbalance, particularly T>Z that completely invalidates the OP %'s (there is no way 50.4% is an indicator of current TvZ balance). A zerg winning GSL is an outlier and even then the top 4 had 3 T's.
|
If u talk about gsl most protoss played rly bad.
even tester vs itr messed up in both games - the first lost it when it was his by letting the 2-3 collossus slip into terran and been sloppy with attack overall , the second cause he screwed up in the opening.
for the other toss in the tournament , gosh i see ladder guys doing better
stats dont lie and if u have any clue from statistics or math you should have know it.
and if not 100% coreect their much better than random opinions.
Atm toss seems the strongest and just a noticed but koreans belive that as well.
also its evident that zerg actually can beat most of the times terran suprising as it sounds.
|
On October 08 2010 23:58 Burban wrote:
if the MMR whatever force the 50% win, because like in chess lets say that terran (white) have a +Xpoints MMR due to racial imbalance, why do we see stats like 55-45 for zerg or 60-40 for toss in some leagues ? Could someone explain that ?
Yea, I don't think people actually read the original post from blizzard. Look at the stats from Gold:
Win % in Gold (accounting for player skill)
61.0% win rate for Protoss when fighting Terran.
61.1% win rate for Protoss when fighting Zerg.
49.5% win rate for Terran when fighting Zerg.
However Blizzard got their stats, they attempted to remove the 50/50 match making phenomenon. How well they did it can be up for debate, but they did and thus I find the numbers meaningful. Though even Blizzard said they base a lot of their balancing on their own experience and not just on random statistics.
I do think there is a separate issue people are raising.Zerg, for example, may be more difficult to play. Meaning an equaled skilled Zerg player has to work harder for his victory. I personally don't have a problem with this. Given the same skill rating, I respect a Zerg player more than the other races, and a Random player the most (as they have to master many more match ups). Just my opinion, but I kinda like having that variety.
|
Since people are saying that there is no conceivable way that they could account for matchmaking, I shall attempt to present a conceivable one.
A = Estimated skill level determined by entire match history
B = Estimated skill level determined by only mirror matches
Delta = B - A
Player Win% for Matchup = Determined by player's matches in that matchup where the differential in MMR skill is ~ Delta
Win% for Matchup = Average of all Player's Win% for Matchup
Profit?
|
when you look at the diamond stats I think the game is alright as it is. that kind of level of balance is hard to achieve. whether this is representative is another thing, but surely all that mid diamond QQ (i'm mid diamond myself) has proven to be false.
|
On October 09 2010 11:24 eloist wrote:
Since people are saying that there is no conceivable way that they could account for matchmaking, I shall attempt to present a conceivable one.
A = Estimated skill level determined by entire match history
B = Estimated skill level determined by only mirror matches
Delta = B - A
Player Win% for Matchup = Determined by player's matches in that matchup where the differential in MMR skill is ~ Delta
Win% for Matchup = Average of all Player's Win% for Matchup
Profit?
You don't even really have to cut out all the games where there's not a big mismatch.
The fact of how MMR systems work is that the MMR (which Blizzard has access to and thus can quantify) determines how likely you are to win based on skill alone vs someone of another MMR.
Lets assume that a 100 difference in MMR means that the higher MMR is likely to win 60% of the time. (Obviously, equal MMR = 50% of the time). And this assumption is only an assumption in numbers - such a value=%wins number does exist.
So lets say you have a 1100mmr vs a 1000 mmr. If the record is not 60-40, that represents an imbalance of some other kind - maybe map based, maybe racial, whatever. The % imbalance it represents for a 50-50 skill matchup is harder to calculate, and the sigma value that describes the certainty of the MMR complicates things further, but nothing in the MMR system makes it impossible to do this.
The only inaccuracy possible is created by the assumption that MMR isn't already affected by racial imbalances, but that inaccuracy is not sufficient to wipe out a 10% discrepancy in win ratios (basically, the Diamond stats say fuck all, but the Plat/Gold/Silver/Bronze stats indicate a real imbalance at the skill level in question).
|
Whats the percent for Zerg vs Zerg?
:D JK
|
On October 08 2010 07:29 Klive5ive wrote:
Win % in Diamond (accounting for player skill)
49.6% win rate for Protoss when fighting Terran.
52.8% win rate for Protoss when fighting Zerg.
49.6% win rate for Terran when fighting Zerg.
Now I'm hoping we could have a brief discussion of what that actually means.
I'm reading it as based upon hidden elo (skill) what percentage of games are won against different races.
So these stats only show the difficulty of the match-ups RELATIVE to each other.
The only reason each race doesn't add up to 100% is because there are not an equal number of players playing each race.
In other words these stats show that ZvP is harder for diamond zergs than ZvT.
But they do not necessarily show that P > Z.
Nor do they necessarily show that Z > T.
Because these stats cannot show what elo(skill) a player would reach if he changed races, since this is obviously pretty impossible.
Thoughts? Am I wrong?
Just wondering where you got this from... looking for updated stats
|
|
|
|
|
|
EPT
