December Aligulac Balance Report

On January 02 2016 17:08 BronzeKnee wrote:
These aren't my stats. Click the link and you'll get all the information regarding sample size and such. The December stats are based on 2683 games.

On January 02 2016 18:04 Topdoller wrote:
More stats which mean nothing, when you consider the maps currently in use.

Also the word "Balance" is misleading in the title of the OP. Shouldn't it say "Win Rate" or something or are you trying to promote your agenda here?

Aligulac is nothing to do with balance, it never has and it never will, especially when the matchmaking system is designed to match player v player

On January 02 2016 18:11 dNa wrote:



those numbers are based on pro games, not ladder games.

On January 02 2016 17:56 Areanon wrote:
When Aligulac begins to count maps in balance? The main problem of today's balance is imbalanced map pool

On July 11 2013 06:39 Grovbolle wrote:
While I won't discuss the statistical/predictional value on the matter of maps or no maps, I can discuss some of the more simple issues:
1: Getting the info on maps played can be tough unless you are talking about the big leagues, let's not get into the problems with different versions of a map and some with forced spawns etc.
2: Backtracking the existing DB would take months since we can't do it in bulks like we did with the event NSM.
3: Number of capable volunteers = 5-10
4: The entire entry system/parsing system would need an overhaul.
5: As far as I recall, we store the rows of matches as sets, not games, meaning that they would needed to be split with an extra attribute to keep track of which games belongs to which set.

I really wish we could have maps, but if you check aligulac.com/db you see that outside kiekaboe, conti, BB, MoP and shellshock, we really doesn't have nearly enough manpower to consistently keep it all updated, nor do we have the prestige of being affilliated with TL to attract them.

On January 02 2016 18:23 Topdoller wrote:


So its nothing to do with Balance.I watched the majority of GSL qualifiers, where a lot of the top players struggled with the new meta and made massive mistakes loosing them the games. The map pool was one of the biggest deciders for the results

"Note that this yields information about metagame balance near the top of the skill ladder, and is not to be confused with (although likely correlated to) actual game balance throughout the whole player population."

On January 02 2016 18:23 Topdoller wrote:


So its nothing to do with Balance.I watched the majority of GSL qualifiers, where a lot of the top players struggled with the new meta and made massive mistakes loosing them the games. The map pool was one of the biggest deciders for the results

"Note that this yields information about metagame balance near the top of the skill ladder, and is not to be confused with (although likely correlated to) actual game balance throughout the whole player population."

On January 02 2016 18:04 Topdoller wrote:
More stats which mean nothing, when you consider the maps currently in use.

Also the word "Balance" is misleading in the title of the OP. Shouldn't it say "Win Rate" or something or are you trying to promote your agenda here?

Aligulac is nothing to do with balance, it never has and it never will, especially when the matchmaking system is designed to match player v player

On January 02 2016 19:06 Inflicted wrote:


It's there if you hover over the dots
987 TvZ
635 PvT
1061 PvZ

On January 02 2016 19:06 zerge wrote:
Please stop using winrates as an equivalent for balance. Balance means equal possibility to win given equal skill level.

On January 02 2016 17:56 Areanon wrote:
When Aligulac begins to count maps in balance? The main problem of today's balance is imbalanced map pool

On January 02 2016 20:26 Ghanburighan wrote:
It's useful to ask how representative the sample size is, i.e., how many people would need to have a bad day to change the result by a percentage point. The answer is 4. If 4 Z players during the month had a bad day, for example a hangover during their game, and they lost 2-0 instead of winning 2-0, the result would be 44%. That's a scary low number.

That basically says that the balance from acceptable short-term winrate of 45% is 8 people, and for perfect balance 45 matches should have gone differently.

On January 02 2016 18:04 Topdoller wrote:
More stats which mean nothing, when you consider the maps currently in use.

Also the word "Balance" is misleading in the title of the OP. Shouldn't it say "Win Rate" or something or are you trying to promote your agenda here?

Aligulac is nothing to do with balance, it never has and it never will, especially when the matchmaking system is designed to match player v player

On January 02 2016 22:46 Arvendilin wrote:

I always thought it was more like this:

Thread gives a reason to whine about Protoss? Awesome great work, Aligulac shows us how the balance currently is.

Thread doesn't give us a reason to whine about Protoss? Meh, but the graphs look really nice and game design guys!

Thread shows Protoss looses a bunch more than winning? Why the fuck are you even posting it here you biased little c***?!

Tho that might be my biased view on it that is heavily tarnished by reddits reaction to these threads =P

On January 02 2016 22:46 Arvendilin wrote:

I always thought it was more like this:

Thread gives a reason to whine about Protoss? Awesome great work, Aligulac shows us how the balance currently is.

Thread doesn't give us a reason to whine about Protoss? Meh, but the graphs look really nice and game design guys!

Thread shows Protoss looses a bunch more than winning? Why the fuck are you even posting it here you biased little c***?!

Tho that might be my biased view on it that is heavily tarnished by reddits reaction to these threads =P

On January 03 2016 01:15 EatingBomber wrote:
The game is balanced for the top of the practical skill ceiling - is it possible a filter could be applied to exclude all non-Korean matches not in the GSL or SSL? Thanks.

On January 03 2016 01:22 StateSC2 wrote:

Sure. Here are the stats for GSL 2016 and SSL 2016.

On January 03 2016 02:13 ApBuLLet wrote:
I agree that calling it a 'balance' report is misleading and implies that the statistics shown are solely a result of balance and no other factors, which is ignorant to assume.

LOTV is a new expansion and people are still figuring it out. The statistics you provided do show that Protoss players aren't doing well in PvZ at the moment, and one possible cause may be balance, but it is ignorant to assume that it is the only reason for them performing poorly.

Maybe part of the reason is that many Protoss players on ladder just deathballed for 5 years and now that style is no longer viable. I don't mean to imply that the players who played this way aren't skilled because I know many of them are, it was just the way Protoss had to play for a while.

But maybe the changes made for LOTV are causing Protoss players to relearn the matchup to an extent. Part of this is that the ravager has given Zerg some early/mid game timings that are being abused a lot on ladder right now. Blizzard has acknowledged that they think these timing attacks are too strong and plan to address it in the next patch by increasing the ravager morph time.

So my point is yes there may be some balance issues and yes these statistics may reflect that, however, it is naive to assume that balance is the ONLY factor driving those statistics because there are many other factors at play.

On January 02 2016 18:11 dNa wrote:



those numbers are based on pro games, not ladder games.

On January 03 2016 02:19 dae wrote:




These stats are not from the ladder.

On January 03 2016 01:38 HellHound wrote:

Every races but mine whines the most my race makes valid complaints.
The reason people don't take this stats seriously is because they don't really show us anything
Note how zerg is by far the highest between JAN 2014 and july 2014
that was the blink era where protoss won 9 tournaments in a row and half of them had pvp finals.
On the otherhand broodlord infestor had around 50% zvp winrate

On January 03 2016 02:17 Elentos wrote:

He even still lives in Korea.

On January 03 2016 01:38 HellHound wrote:

Every races but mine whines the most my race makes valid complaints.
The reason people don't take this stats seriously is because they don't really show us anything
Note how zerg is by far the highest between JAN 2014 and july 2014
that was the blink era where protoss won 9 tournaments in a row and half of them had pvp finals.
On the otherhand broodlord infestor had around 50% zvp winrate

On January 03 2016 03:48 TheWinks wrote:
If we had blindly followed aligulac win rates the widow mine revert would have never happened. Aligulac's "balance" page has always been a poor source of information.

On January 02 2016 19:15 keglu wrote:


We don't have tools to measure skill level.

On January 03 2016 03:49 Lexender wrote:


The blink era was entirely the fault of awful maps and everybody knew that and thats how it was resolved, I have no idea what you are talking about

On January 03 2016 03:49 Lexender wrote:


The blink era was entirely the fault of awful maps and everybody knew that and thats how it was resolved, I have no idea what you are talking about

On January 03 2016 03:48 BronzeKnee wrote:


Why would we blindly follow the win rates? I don't see anyone arguing for that. They are a source of information to be considered.

But not ignored. Therefore we should neither follow them blindly, nor totally discard them. The fact that PvZ is at it's worst point in terms of win rates ever should be very concerning and is cause for some kind of change.

On January 03 2016 04:54 Lexender wrote:


I remember when ByuN was being number 1 terran, in the middle of HotS

On January 03 2016 06:54 Nebuchad wrote:


We know how it works. When people play a lot of online tournaments, they end up advancing their rating way more (and way easier) than the people who play only GSL and the occasional week-end tournament, even though they are playing weaker opposition. You have demonstrated that the ranking is flawed, but that's completely different from collecting results and calculating a win percentage based on them.

What is described as "garbage" is simply things that contradict the narrative you're pulling for.

On January 03 2016 07:12 TheWinks wrote:

The two things you're describing are linked. If a top or near top level Korean goes into an online tournament and wipes the floor with a bunch of foreigners he gets a bigger boost in aligulac rating that he really should AND all those games get added into the win rate graph despite their lower quality.

On January 03 2016 08:22 Big J wrote:

and the underlying assumption unless proven differently (statistically or deductively, not by giving a piece of evidence) is:
such a player can be of any race, so the winrates get influenced by it in the same ways

also you gotta prove that the samplesize is small enough for it mattering to begin with. time for some math if you want to support your hypothesis

On January 03 2016 09:17 Athenau wrote:
In other words, by symmetry there's no reason to expect that a much-lower skilled Z plays a higher skilled P more often than the other way around.

On January 03 2016 08:43 TheWinks wrote:

There is clearly not an equal race distribution just by looking at the balance report itself.

TvZ: 987
PvT: 635
PvZ: 1061

Why would we assume what you're saying? Even if low quality games did mostly cancel each other out, why include them? It's unnecessary noise. Also, your second sentence doesn't make any sense.

There's no reason to make this assumption though. We clearly have a race population mismatch.

On January 03 2016 09:30 Big J wrote:
that's why comments such as yours trigger me to respond in such a way. you try to diminish numbers without doing actual work.

On January 03 2016 09:51 Athenau wrote:

Relative size doesn't matter. Even if you have, for example, a population of 100 zerg players and only 10 protoss players, the number of ways to match a top 10% zerg player with a bottom 10% protoss player is the same as the other way around. The problem is symmetric.

David Kim wrote:
It’s the only major tournament going on right now, and it represents a sample size that is too small to draw any broad conclusions from, but Protoss players have lost at a noticeable clip in that tournament. As of this writing, they’ve recorded only 11 wins in 35 non-mirror matchups.

On January 03 2016 09:51 Athenau wrote:

Relative size doesn't matter. Even if you have, for example, a population of 100 zerg players and only 10 protoss players, the number of ways to match a top 10% zerg player with a bottom 10% protoss player is the same as the other way around. The problem is symmetric.

On January 03 2016 11:50 Lexender wrote:


No because its not a randomized experiment where you check all variants, if there is 100 zerg players and 10 protoss players, sure at some point a top zerg and a top protoss will meet, but until that point the protoss players may have defeated 10 bottom zerg players while the zerg player may have played ten mirrors wich will skew balance, wich is why not only non-mirrors have to be checked but mirrors too.

On January 03 2016 11:59 cheekymonkey wrote:


This is exactly why it makes sense to weight the win/loss statistics according to the players respective ELO rank parameters. So that if a zerg player has a 60% chance of winning against a lower level protoss, a win will be weighted as some number < 1, and conversely the protoss winning will be weighted by some number > 1.

However, given an imbalance meta, I suspect that in time as the ELO rankings will adjust and stabilize, it will stop representing the relative skill levels between players, and rather reflect a combination of skill level and current racial advantages/disadvantages. So in time, this weighted calculation might stabilize around even racial winrates. I would appreciate someone well-versed in statistics to chime in on this.

On January 03 2016 13:06 dainbramage wrote:


What you're asking for is the bottom graph in the Aligulac report. But you run into the same problem here. Racial balance influences players' Elos, much like player skill influences overall winrates, so it'll always trend towards 0. The only thing it might help you with is determining how a change affects a matchup. E.g. if the map pool changes and you see zerg shoot up, you know the new pool is more zerg favourable than the last one - but only as a comparison to the previous pool, not that it's necessarily 'good'. And you'll see it fall back to even as Elos adjust.

EDIT: Unrelated, but whenever these threads pop up, you'll always have one party saying that the statistics are definitive proof, and probably using p-values from binomial tests with invalid assumptions. Then you'll have another saying that they should just be completely ignored because of selection bias and they're not independent and players have differing skills. Funny how the people treating them as gospel play the race doing poorly, and those claiming it's invalid are the ones doing well.

On January 03 2016 13:06 dainbramage wrote:


EDIT: Unrelated, but whenever these threads pop up, you'll always have one party saying that the statistics are definitive proof, and probably using p-values from binomial tests with invalid assumptions. Then you'll have another saying that they should just be completely ignored because of selection bias and they're not independent and players have differing skills. Funny how the people treating them as gospel play the race doing poorly, and those claiming it's invalid are the ones doing well.

On January 03 2016 13:06 dainbramage wrote:


What you're asking for is the bottom graph in the Aligulac report. But you run into the same problem here. Racial balance influences players' Elos, much like player skill influences overall winrates, so it'll always trend towards 0. The only thing it might help you with is determining how a change affects a matchup. E.g. if the map pool changes and you see zerg shoot up, you know the new pool is more zerg favourable than the last one - but only as a comparison to the previous pool, not that it's necessarily 'good'. And you'll see it fall back to even as Elos adjust.

EDIT: Unrelated, but whenever these threads pop up, you'll always have one party saying that the statistics are definitive proof, and probably using p-values from binomial tests with invalid assumptions. Then you'll have another saying that they should just be completely ignored because of selection bias and they're not independent and players have differing skills. Funny how the people treating them as gospel play the race doing poorly, and those claiming it's invalid are the ones doing well.

On January 02 2016 20:26 Ghanburighan wrote:
This is somewhat disingenuous by Bronzeknee as such stats have been discussed in the designated dalance discussion thread already and been found wanting. Bronzeknee was part of these discussions so he should know better.

That's not to say that there isn't an early indication of a problem in PvZ.

On January 02 2016 16:59 BronzeKnee wrote:

Protoss is pretty much in free fall, especially in PvZ (haven't had a win rate that imbalanced for three years, and PvZ is more imbalanced than ever), and yet nerfs are planned for Pylon Overcharge and the Adept... interesting.

On January 03 2016 13:06 dainbramage wrote:


What you're asking for is the bottom graph in the Aligulac report. But you run into the same problem here. Racial balance influences players' Elos, much like player skill influences overall winrates, so it'll always trend towards 0. The only thing it might help you with is determining how a change affects a matchup. E.g. if the map pool changes and you see zerg shoot up, you know the new pool is more zerg favourable than the last one - but only as a comparison to the previous pool, not that it's necessarily 'good'. And you'll see it fall back to even as Elos adjust.

On January 03 2016 20:56 Sapphire.lux wrote:

It's also interesting how you conveniently forget to mention that there are 2 significant Zerg nerfs on the horizon: Zergling attack speed upgrade bonus decreased from 40% to 30% and Viper spell damage reduced from 90 to 60.

The PO i'd call more of a change rather then nerf, as it get stronger in some situations and not so much in others.