|
I don't remember you showing me winrates where terrans did badly in korea vs protoss at that time. All I remember is the numbers I pulled, 136-93 for terran.
I don't know where you are getting those numbers from.
Win/rates for GSL in 2012 (the five seasons)
P 50.0% 49.2% 51.1% 46.9% 48.3%
T 50.0% 50.8% 48.9% 53.1% 51.7%
WCS Korea season 1. PvT = 53.3%
T 48.6% 41.2% 41.4% 39.8% 36.0%
Z 51.4% 58.8% 58.6% 60.2% 64.0%
WCS Korea season 1. TvZ = 47.3%
Source: Aliguac.
Includes Code S, Code A, qualifications.
Regarding the adjusted Aliguac win/rates. Terrans averaged a win/rate of of 47% vs protoss from July 2012 to February 2013. While this isn't signifciant, combined with terrans doing worse vs zerg than protoss, terrrans fell off the competitive scene and zerg gained. Protoss on the other hand didn't suffer. And that's the most important part of the story.
Given that, we have en explanation for why terrans have a lower adjusted win/rate post-HOTS --> Most people simpy weren't aware that the drop-out of terran players happened in 2012, and not just over 6-8 months ago.
|
On September 14 2014 04:07 Hider wrote:Show nested quote +On September 14 2014 04:04 Nebuchad wrote:On September 14 2014 04:00 Hider wrote:On September 14 2014 03:56 Nebuchad wrote:On September 14 2014 02:36 Hider wrote:
In page 1079 (or 1080) I discussed some of the possible explanation for why terran was preceived (amongst the general community) to be in a better positon than the numbers suggested by the adjusted win/rates.
In the end it comes down to terran suffering more than protoss during the patch-zerg, which wasn't a very well-known story.
I'm still contesting that in the same way I did at page 1079-1080 :p Well you made a comment that terrans did fine in Korea. Then I looked at win/rates and saw that wasn't true. Terrans weren't doing well vs protoss. Instead, the perception is probably a consquence of race distribution in code S/WCS Korea, which I in my post above - argued is a really bad indicator of balance. I don't remember you showing me winrates where terrans did badly in korea vs protoss at that time. All I remember is the numbers I pulled, 136-93 for terran. I don't know where you are getting those numbers from. By using data from both all of GSL and WCS Korea (code A, code S, qualifications), I made this graph.
I think it makes a lot more sense to put WCS Korea in the preceding time period. The period before patchzerg era had protoss doing the best, with the MC-Seed final and an overall better performance in code S even in the season Mvp won. If you put WCS Korea in that period, you have a continuation of trends.
|
On September 14 2014 04:15 Nebuchad wrote:Show nested quote +On September 14 2014 04:07 Hider wrote:On September 14 2014 04:04 Nebuchad wrote:On September 14 2014 04:00 Hider wrote:On September 14 2014 03:56 Nebuchad wrote:On September 14 2014 02:36 Hider wrote:
In page 1079 (or 1080) I discussed some of the possible explanation for why terran was preceived (amongst the general community) to be in a better positon than the numbers suggested by the adjusted win/rates.
In the end it comes down to terran suffering more than protoss during the patch-zerg, which wasn't a very well-known story.
I'm still contesting that in the same way I did at page 1079-1080 :p Well you made a comment that terrans did fine in Korea. Then I looked at win/rates and saw that wasn't true. Terrans weren't doing well vs protoss. Instead, the perception is probably a consquence of race distribution in code S/WCS Korea, which I in my post above - argued is a really bad indicator of balance. I don't remember you showing me winrates where terrans did badly in korea vs protoss at that time. All I remember is the numbers I pulled, 136-93 for terran. I don't know where you are getting those numbers from. By using data from both all of GSL and WCS Korea (code A, code S, qualifications), I made this graph. I think it makes a lot more sense to put WCS Korea in the preceding time period. The period before patchzerg era had protoss doing the best, with the MC-Seed final and an overall better performance in code S even in the season Mvp won. If you put WCS Korea in that period, you have a continuation of trends.
I removed the graph. It was dumb and confusing. I guess the point here isn't whether PvT was imbalanced in Korea or not (seems pretty balanced from the numbers), but rather as an attempt to explain why competitive terrans dropped off when protosses didn't. The altnerative here is to assume that there is no balance-related reason for the drop-off and that it can already be explained by "structural changes". It's a bit hard to find any clear-cut evidence here, but the story that "terrans suffered more than protosses makes during patchzerg"-era makes the most sense to me.
Given how it seems that Aliguac adjusted win/rates today matches the performance of Korean terrans, I have trouble seeing a supprt for the "skill ceiling"-argument. Thus I believe the adjusted win/rates of Aliguac today are an useful indicator of balance at the highest level (better quality than the low sample size from GSL).
|
On September 14 2014 04:11 Hider wrote:Show nested quote +I don't remember you showing me winrates where terrans did badly in korea vs protoss at that time. All I remember is the numbers I pulled, 136-93 for terran. I don't know where you are getting those numbers from. Win/rates for GSL in 2012 (the five seasons) P 50.0% 49.2% 51.1% 46.9% 48.3% T 50.0% 50.8% 48.9% 53.1% 51.7% WCS season 1. PvT = 53.3% Source: Aliguac. Includes Code S, Code A, qualifications. Regarding the adjusted Aliguac win/rates. Terrans averaged a win/rate of of 47% vs protoss from July 2012 to February 2013. While this isn't signifciant, combined with terrans doing worse vs zerg than protoss, terrrans fell off the competitive scene and zerg gained. Protoss on the other hand didn't suffer. And that's the most important part of the story. Given that, we have en explanation for why terrans have a lower adjusted win/rate post-HOTS --> Most people simpy weren't aware that the drop-out of terran players happened in 2012, and not just over 6-8 months ago.
My numbers are from code A and code S in the last three seasons of WoL.
Not taking WCS Korea cause I believe it fits better with the preceding trend of protoss being the leading race.
Not taking the qualifiers, at first because I was too lazy to get them, and now because even if they impact the trends so much, it fits better in my argument than yours.
But a main question is, why do you feel like you need a PvT imbalance to explain that dropout of players? I think it's pretty easy to establish that, while number of players have an effect on winrates, winrates don't always have an effect on number of players. The fact that there was more protoss players than terran players in aligulac in october 2011 should be enough to prove that. Your overall argument doesn't suffer.
|
But why do you feel like you need a PvT imbalance to explain that dropout of players? I think it's pretty easy to establish that, while number of players have an effect on winrates, winrates don't always have an effect on number of players. The fact that there was more protoss players than terran players in aligulac in october 2011 should be enough to prove that. Your overall argument doesn't suffer.
It would probably be too much to say imbalance of a 53% average (though it was very consistent). But you definitely have a point: While adjusted win/rates in Aliguac can be useful to determine balance, they can not explain a drop-off. Only naked win/rates can exchange a drop-off due to balance and anything else could be due to a change in sample size quality or structural reasons.
If we assume sample size quality is unchanged throughout the period, I think I can come up with (a consistent) estiamte for structural changes.
|
On September 14 2014 01:40 Hider wrote:Show nested quote +It would be good to call it something other than a win rate, say 'balance indicator' or 'strength indicator'. Otherwise you will always get flack for it not actually being a true win rate. I called it adjusted win/rates, which is exactly what it is - Not sure whether that's worse or better than balance-indicator.
Technically you could call everything that has win rate as input an adjusted win rate. But that does provide much information. I agree that my terms where also not very informative, but they also were not misleading.
Typically an adjusted rate of something would be because you take less than the total sample set, correct other problems with the data, ... This is not your argument though, is it? You are basically saying that correcting for unequal distributions via (A+1)/(B+1) gives a better indicator of historic performance as measured by widely accepted perception of race strength? That is, the justification is empirical/statistical through 'backtesting' rather than theoretical? (Apologies if someone posted conclusive theoretical arguments for why an overrepresented race should have its win rates adjusted downwards.)
|
So I looked a bit of the numbers and ran some correlation calculations. While there generally is a postivie correlation between an increase in win/rates and representation %, there is one big problem with my thesis.
For some reason the representation of terrans declined in 2011 despite win/rates being above 50% for most of the year.
Thus, it seems there is a consistency problems when arguing that the only reason terran representation declined was becasue the race is was weak (measured by the naked win/rate).
Instead, it seems more likely that it's a combination of structural factors and a change in balance (with the latter weighting less).
Given that, I decided to revisit my philosophy on balance a bit, and I developed an alternative to approach to adjusting win/rates:
- If one race has a a win/rate less than +/- 1.5% away from 50% 3 months in a row, it is assumed that the race has succesfully readjusted
- Going forward, the representation % for those 3 years will be applied.
- Adjusted win/rates will then be calculated as: Naked XvY win/rates * (1+A-B)
- A = Actual representation % - targetted representation of race X
- B = Actual representation % - targetted representation of race Y
So an increase in targetted representation % increases the win/rates of the race.
In the below graph, it can be seen that terran still struggled (more than toss) durng the patch-zerg era, and we also see no support at all for the Queen range buff (TvZ averaged 50% the 3 months prior).
In early HOTS terran was imba both matchups. Widow Mine nerf was, however totally unneeded as win/rates averaged close to 50% in the months prior to the nerf.
It's also quite obvious that protoss was OP for a long period of HOTS as win/rates consistnetly ranged between 52%-56% A couple of months ago - it seemed that terrans figured out the matchup, but then TvZ started to get imbalanced in favor of zerg.
As of August, terran seems a bit too strong in both matchups (but still within the +/- 5% threshold). We will probably need an additonal month or two in order to come up with a better assesment.
PvZ seems pretty damned balanced.
![[image loading]](http://i61.tinypic.com/2hrg3zc.png)
|
I just hope blizzard lays low on the balance pathes now until LotV
|
On September 14 2014 22:24 cheekymonkey wrote:
I just hope blizzard lays low on the balance pathes now until LotV
So do I. If they are not willing to make drastic changes to the game desing before the next expansion, better keep the game how it is for now. I'd say it's almost impossible to significantly improve the quality of this game by only tweaking unit numbers, things need to be redesigned.
|
On September 14 2014 22:24 cheekymonkey wrote:
I just hope blizzard lays low on the balance pathes now until LotV
Still looking for that Siege tank + damage vs shield buff - balance aside - I would like to see how bio + tank + full mech could work vs toss. If toss wants a compensation I think we could nerf scv-pulls somehow. Perhaps reintroduce Ghost energy or perhaps make it easier to get Templars out.
|
On September 14 2014 22:53 Hider wrote:Show nested quote +On September 14 2014 22:24 cheekymonkey wrote:
I just hope blizzard lays low on the balance pathes now until LotV Still looking for that Siege tank + damage vs shield buff - balance aside - I would like to see how bio + tank + full mech could work vs toss. If toss wants a compensation I think we could nerf scv-pulls somehow. Perhaps reintroduce Ghost energy or perhaps make it easier to get Templars out.
The removal of Moebius Reactor had no effect on SCV pulls since Moebius Reactor was not searched before and Ghosts would get 75 energy walking across the map. (Just saying.)
|
On September 14 2014 23:05 TheDwf wrote:Show nested quote +On September 14 2014 22:53 Hider wrote:On September 14 2014 22:24 cheekymonkey wrote:
I just hope blizzard lays low on the balance pathes now until LotV Still looking for that Siege tank + damage vs shield buff - balance aside - I would like to see how bio + tank + full mech could work vs toss. If toss wants a compensation I think we could nerf scv-pulls somehow. Perhaps reintroduce Ghost energy or perhaps make it easier to get Templars out. The removal of Moebius Reactor had no effect on SCV pulls since Moebius Reactor was not searched before and Ghosts would get 75 energy walking across the map. (Just saying.)
Would they? As I remember it, it takes around 1.4 seconds to gain 1 energy (?) Thus that's like 40 seconds. Probably takes around 30 seconds to walk across the map, but yeh - it's probably almost irrelevant.
Better to make templar tech cheaper as it perhaps could buff HT/Zealot which is more entertaining than Collosus-play.
|
On September 14 2014 23:15 Hider wrote:Show nested quote +On September 14 2014 23:05 TheDwf wrote:On September 14 2014 22:53 Hider wrote:On September 14 2014 22:24 cheekymonkey wrote:
I just hope blizzard lays low on the balance pathes now until LotV Still looking for that Siege tank + damage vs shield buff - balance aside - I would like to see how bio + tank + full mech could work vs toss. If toss wants a compensation I think we could nerf scv-pulls somehow. Perhaps reintroduce Ghost energy or perhaps make it easier to get Templars out. The removal of Moebius Reactor had no effect on SCV pulls since Moebius Reactor was not searched before and Ghosts would get 75 energy walking across the map. (Just saying.) Would they? As I remember it, it takes around 1.4 seconds to gain 1 energy (?) Thus that's like 40 seconds. Probably takes around 30 seconds to walk across the map, but yeh - it's probably almost irrelevant. Better to make templar tech cheaper as it perhaps could buff HT/Zealot which is more entertaining than Collosus-play.
25 energy = 45 seconds, which on average is the time needed for a Ghost to walk from the rax to the targeted location.
|
Sorry about any spelling/gramatical mistakes and I'm not sure if has been discussed before:
After watching a ton of TvZ games and more recently Snute vs MMA, I see that 2 rax openers is pretty strong, unless the zerg scouts it with luck or if he doesn't go hatch first, it has a increadible success rate. It's very difficult to hold without taking sigificent losses and the terran seem to be in front no matter what as long as he doesn't make any huge mistakes. Maru is showing this aswell, with him 2 raxing most of his games. How I see it a cheese have become a more standard opening in TvZ, which annoys me a bit, a cheese should be pulled out to suprise your oppnent, not a standard opener that you will feel confident succeding with. Does proxy two rax need a slight nerf or am I completely wrong and simply shouting "hurr durr 2rax op", please share your opinions.
I would love to see what kind of win rates 2 raxing in pro matches have.
|
On September 15 2014 00:40 Chillidawg wrote:
Sorry about any spelling/gramatical mistakes and I'm not sure if has been discussed before:
After watching a ton of TvZ games and more recently Snute vs MMA, I see that 2 rax openers is pretty strong, unless the zerg scouts it with luck or if he doesn't go hatch first, it has a increadible success rate. It's very difficult to hold without taking sigificent losses and the terran seem to be in front no matter what as long as he doesn't make any huge mistakes. Maru is showing this aswell, with him 2 raxing most of his games. How I see it a cheese have become a more standard opening in TvZ, which annoys me a bit, a cheese should be pulled out to suprise your oppnent, not a standard opener that you will feel confident succeding with. Does proxy two rax need a slight nerf or am I completely wrong and simply shouting "hurr durr 2rax op", please share your opinions.
I would love to see what kind of win rates 2 raxing in pro matches have.
Its a wax wane thing. Heres how it happens.
Terrans 2rax a lot, leading to a lot of wins. Zerg start playing safer, some by going pool first while others by keeping 2 drones out front as anti-bunkers.
2rax stops winning very much, and most 2rax starts losing. Terran starts playing more "standard/macro" oriented. Zerg stops playing safe, and we go back to the 2rax phase.
|
On September 15 2014 00:40 Chillidawg wrote:
Sorry about any spelling/gramatical mistakes and I'm not sure if has been discussed before:
After watching a ton of TvZ games and more recently Snute vs MMA, I see that 2 rax openers is pretty strong, unless the zerg scouts it with luck or if he doesn't go hatch first, it has a increadible success rate. It's very difficult to hold without taking sigificent losses and the terran seem to be in front no matter what as long as he doesn't make any huge mistakes. Maru is showing this aswell, with him 2 raxing most of his games. How I see it a cheese have become a more standard opening in TvZ, which annoys me a bit, a cheese should be pulled out to suprise your oppnent, not a standard opener that you will feel confident succeding with. Does proxy two rax need a slight nerf or am I completely wrong and simply shouting "hurr durr 2rax op", please share your opinions.
I would love to see what kind of win rates 2 raxing in pro matches have.
I disagree on the 'íts a phase'' comments. 2 rax has always been a very strong opening. Maru opens 2 rax almost every single game because he knows the worst case scenario is being slightly behind, Zerg has to go low eco to stop it so it evens out in the end. The only scenario where 2 rax is bad is when zerg opens hatch first and holds it with minimal drone losses, which almost never happens against competent terrans. Cheese should be bad, it should only punish greedy/bad players, otherwise its very shitty for viewers and players.
|
Besides, often a 2rax opening leads into a longer game too. It's not like it was before, it's a more common opening now because of this.
|
Hellbat change has made 2 rax much stronger I think because the transition is a LOT faster when you're on low econ. The way we see these 2 rax winning is not always outright.. but rather they do crippling economic damage so the Zerg doesn't have the option to go roaches (or be even more behind). Hellbats being more accessible has changed things.
|
On September 15 2014 01:25 DinoMight wrote:
Hellbat change has made 2 rax much stronger I think because the transition is a LOT faster when you're on low econ. The way we see these 2 rax winning is not always outright.. but rather they do crippling economic damage so the Zerg doesn't have the option to go roaches (or be even more behind). Hellbats being more accessible has changed things.
And in case of an attempt to baneling bust in response, tanks have free siege mode and mines got buffed. So why not 2 rax most times, as long as you're not getting blind countered?
I'm having a hard time seeing the downside to this cheese.
What's the last case of this getting defended (not via blind counter) and causing the terran to lose the game, or at least to be far behind?
|
I don't know if I'd consider 2-rax cheese. It typically leads to a pretty even situation for both parties assuming equal levels of play and with the way each race's follow-up works it almost always transitions into a "normal"-looking game. I'd consider it more of a committed pressure build.
It hits super early, yeah, but it isn't cheesy in how it plays outside of a few map-specific bunker positions.
|
|
|
|
|
|
EPT
