Fact check: Zergs win everything

General Discussion
1

You’ve no doubt heard some red-eyed whiner crying in some obscure corner of the forums, balling his eyes out, saying that “Zerg is winning EVERYTHING!”, and “They are going to win blizzcon, too!” and of course these are always followed by “The infestor is so OP” and “Zerg late game is unbeatable” and “queens need to be nerfed.”

While their emotive performances are deserving of an Emmy nomination, the real world functions on this little thing called reality and we measure that with evidence. While this may feel like a sucker-punch to the gut of the whiners, I feel it is my duty as a rational human being to educate the forums on the reality of the game’s state. Those glorious facts are:

  1. There have been 2002 events in 2019 according to Aligulac. That’s a whole 'lotta tournaments that are all conveniently ignored in favor of the highest profile events that are usually dominated by Serral. There is no surprise here why the whiners focus on these events. They look no further than to find something that confirms their assumptions.
  1. Each participant on average should have 1/totalNumberOfPlayers odds of winning the tournament. In a round of 32 tournament, each player on average will have 1/32=3.125% chance of winning the whole tournament.
  1. The odds for a race to win a tournament is a product of the participants. If each player has a 1/32 odds to win, then if there are 10 zergs, zerg collectively should have 10/32 odds to win. Not rocket science, just basic math.
  1. The tournaments on average were 25.7% Terran, 31% Protoss, 42.9% zerg.
  1. 575 of these tournaments were won by Terrans (29%). 868 were won by Zergs (43%). 569 were won by Protoss (28%).
  1. Zergs win 42.9% of tournaments while making up 43% of tournaments. That is perfectly balanced. That means every zerg in a round of 32 tournament will have a 1/32=3.125% chance to win.
  1. Protoss win 29% when they ought to win 31%. Protoss is underpowered.
  1. Terrans win 29% of tournaments when they should only win 25.7. Terran is overpowered, quite probably vs protoss. In a round of 32 tournament, there would be an average of 8.224 terran players who collectively have a 29% chance to win the whole tournament. This means each player has a 3.5% chance to win the tournament on average, when a fair win-rate would be 1/32=3.125%.
  1. We can punch the Terran data into [https://www.stattrek.com/online-calculator/binomial.aspx] to calculate how confident we are in Terran being OP. This could happen by chance, so we need to know what the odds are that it did happen by chance. This yields a 0.011% chance that this occurred by chance. There is a 99.989% chance Terran is overpowered.

Conclusion

Zerg is doing exactly as it ought to when it comes to winning tournaments, for the first time since 2015, and Terran is overpowered vs Protoss with an exceedingly high confidence. If you listened to the forums, Protoss and Zerg needed to be OBLITERATED with the nerf-hammer, but reality dictates otherwise. In fact, Protoss were just hit with the nerf-hammer and are being hit with another nerf-NUKE in the end of the year changes.

As you can see, there is little reason to believe that Blizzard balances the game based on tournament wins. I think they listen entirely to what the whiny players say and they said NUKE PROTOSS PLZ and that’s what blizzard is doing. I for one would like to see the balance team use logic & evidence to balance the game.

14 Likes
2

Well, but you forgot to take playerskill in your calculation. F.e. a serral should have a better chance to win a ro32, than 1/32.

Also, with that skill factor in mind, weekly tournaments arent always usable, as a statistic. If f.e. any pro is in a tournament, while all other players are m2 or below, its clear, that the pro will smash through the bracket, no matter which race he plays.

But yeah, on pro lvl, its mostly serral only and then these whiners freak out, that zerg wins everything. And spoiler - serral will win blizzcon again. Besides dark, theres noone, that can beat him atm. But whiners will go nuts, thats for sure.

8 Likes
3

Minor tournaments carry no weight. One or two pros are present for each and incidentally one of those always wins.

What a coincidence.

First thing you are taught in university when working with data is how to spot dishonest use of them.

You are trying to push data that read: pro gamer wins 95% of the time to mean: “pro gamers race matters”.

It does not. Compare tournaments where players of similar level play. Then you gain relevant data and see it goes against your agenda.

Try better next time.

18 Likes
4

I love people try to argument using data. You have mi respect. Sadly, I cant agree with you (OP) in your conclusions.

I have experience analyzin videogame data. When I look for balance analysis, the first parameter to bring is the population distribution, not the winrates. Why?

  1. In PvP ranked games, winrates tend to be close to 50%. If your race, champion, or whatever the tool you use to win in the game is underpowered, the system will give you a lower ranking, then you will play with people with less skill than you but with better tools. The winrates are going to be 50%, even thinking you are using the UP tools.

  2. Its a FACT that people use always the easiest and fastest way to do something. In videogames you can see how ALWAYS the OP tool is the most populated. People are really intelligent (we are all humans) and (most of them) experiment how being in a 50% winrate, they get better results (higher ranking) with the OP tool.

I did this study some time ago in Sc2:

Not only Zerg is the most populated in the higher leagues and pro level, even thinking is the least played race in general in the game. Not only, furthemore, the population distribution depicts how Z raise their population with each higher league.

If you could find a book about “Balance in videogames”, Zerg could be a perfect example of OP. “Easy and fast to climb up in the ladder, even until pro-scene”.

But OK, forget all the above and go to analyze tournament winrates vs population, even thinking that a bigger population of Z in tournaments is only due only to “Better players simply like more Zerg race, due to aesthetical or another reasons”.

The problem with this analysis is that (how SouthWinD said) is that your assumption of “each player has the same probability to win the tournament” is false. If there are 29 pro, and 3 “ultra pro” in the tournaments, probably the 29 poor really good players have a chance close to 0% to win the tournament, meanwhile the another 3 have a chance close to 33%. You can not oversimplificate this if you want to reach real conclusions. Probably you have to assign to each player in the tournament a chance depending on their history results, and normalize this chances (always the sum of probabilities must be 1 in any case). And even then, this is not enough. What do happen with new players? or if a player changed his level of skill incredibly since the last data recorded about him? Anycase, could be a good point to start. But your maths cant be a first aproximation because your hypothesis are not true.

When I see this kind of analysis, then I could believe the conclusion of the data analysis. Until then we only see zerg winning and winning. People always say “It is due to serral”, but only thing we can really do is ask us:

Would serral be so dominant if he would play another race? All we know Serral is the best in the world, but would be so so big the distance between him and the other top 10 with another races? ummmmmmmmmmmmmm. And in the mouth of other Zerg defenders: “only Dark or Rogue can beat him”. What a coincidence, the others ultra pro Zergs.

When you start to see dominance of a race, and the defenders of this situation always talk about it with the arguments “they are really good players”, you can only think, wow and the others ultra pros of the other races are not so good, why?. This is so dangerous and tell us that something is not working how it should be.

14 Likes
5

Balance whining drops to 0%

2 Likes
6

the problem is, you cant objectively rate the skill of a player. theres nothing like f.e. “serral is lvl 99, heromarine is lvl 85, so serral is clearly the better player, nothing to do with balance” etc.

skill is always subjective, so its really hard, to really say, if one player is winning because his race is slightly op, or because hes just the better player. call me a fanboi, but serral is the only one (and MAAAAAABE dark atm) where i would say, that he is objectively better than anyone else, no matter the race. but again, this is my opinion, so in theory, it cant be that objectively… you understand me ;D

the problem is, there is no objective way to messure skill. mmr is the closest thing to objective we have, but even mmr doesnt really matter much. you cant say, that someone is better than you, just because he has like 200 mmr more and that he will win everytime against you.

so with skill not being able to messure, we have no way to separate the balance from the playerskill. this is what makes balance discussions so pathetic. one side will say, that he won because of balance, the other side, that he won because of skill. whos right? who knows.

regarding to that “easiest way to win” stuff… if infestor would be THAT op right now, everyone would use them, since it would be the easiest way to win. instead, we see dark and serral rarely abuse them. the only one, who really abused them lately was rogue at the gsl 3 finals against trap. so if your theory is right, that players always choose the easiest way, and if the forum is right by saying infestor are op… why do these two dont abuse infestor on a regular basis?

7

Thats the dumbest way to calculate it, your logic is so flawed its laughable

1 Like
8

I am not sure. Aligulac has some interesting prediction models about the performance of players. Since I dont know the maths behind them I can not say if It is possible or not (obviously always the models are going to be aproximations, not the perfect reality).

The individual perception of the balance is subjetive, It is true. The ultra pro data is not big enough according to the statistics laws to get clonclusions. This is why I always use the population behaviour to get some conclusions. We have empirical knowledge about the social behaviour, about how people make decisions. It is a solid base to start with.

I agree that infestor is not the problem here, and how you say empirical data show this. Hovewer, I support the idea of Zerg being a little bit favored. The reasons…I would have to write a long post about that, and infestor is only ONE OF MANY of the problem ONLY in late game. Perhaps I can write some day this post. Analizing every aspect of the race, but probably I need more time playing and watching pro games to have the complete analysis, and I dont have this time. This is the task that balance team have to do and is paid for. I have enough with my Robotics PhD.

Only think I can use is the data we have about population and the solid base of human behaviour to get results objetively. The only arguments against I read are: “this is not true” or “Zerg are better players”. Then, until I discover something better, this conclusion is an objetive truth for me.

1 Like
9

true, aligulac has interesting statistics and most of the time, the win prediction from aligulac is right, but if you think about it, its also something, that doesnt really say anything.

if i take f.e. player A vs player B, player A has a 70% winchance against player B regarding to aligulac. this calculation can be right for both cases, player A win or player B. you could only check these numbers, if these two players play a lot of matches against each other, like 100+. even with a 70% winchance, its absolutely doable for player B to win, without negating the calculations.

tbh, to balance the game is a really hard job. almost every whiner in this forum is acting like its just as ez as breathing, but its everything but that. its really hard to judge, if f.e. the infestor is op, or if the players are just better than their opponent. like you said, the common sense would be to observe a bigger population. you cant observe every laddergame ever played. so to observe the thousands of “normal” players is theoreticly the best way for judging the abuse of a specific unit/strat, but in reality, its not doable, cuz of the huge amount of data/time needed.

so that leaves the blizz team with the option to analyze pro games. at that stage, as i allrdy said, infestors arent that abused, so in my point of view, its absolute understandable, if blizz says thinks like the infestor isnt op. is that the case? i dont know.

what you see in this forum, is that people in here leave out the playerskill-factor. yes, in the hands of f.e. a serral, infestors look REALLY terrifying, but tbh, almost every unit looks that way in his hands. that doesnt have to mean, that normal laddergames are influenced by it.

its normal human behavior, that you search mistakes and excuses for your failure. the easy way is to blame it on someone else. the hard and true way is to look at yourself. so i dont really blame guys like playa, platys etc for whining 24/7, since i believe, in their inner core, they know, that theyre the reason why they suck, not the balance.

10

2019

Rogue won (GSL Season 3)
and Dark (GSL S2, GSL Super Tour)
and Serral (1 WCS, GSL v World, HSC vs TY)
Reynor won (WCS)
even SOO won IEM

ZvZ of the most common matchups

Terran won (WESG) - Yes this can add to Terran wins this year
Terran won GSL Season 1 that’s it, end of Terran wins.

Any non-Zerg winning lately? But claim how T is actually OP hahah this clown (the OP) still goes on to amaze.

Now it’s not like last year only Serral. otherwise I agree but guess what we were talking last year? Maru won 4 GSLs Terran Domination OMAGAASH - those who claimed it were bigger clowns than whine about Z wins cause look at the above

7 Likes
11

Oh you just made a mortal enemy in Batz, lol.

6 Likes
12

I have to disagree somewhat about people choosing the easy way out when it comes to race selection.

If people took the easy way out they would just switch races instead of getting emotional about balance. I don’t see a lot of people jumping ship when the race they play gets nerfed. I also don’t see them joining the dark side and playing the race that gets buffed.

I think people new to the game just choose the race they think looks the coolest to them. I like the idea of playing a hive mind race. I’ve liked Zerg since the original Starcraft.

People are loyal to thier race regardless of buffs and nerfs and tournament wins. It’s easier to switch to the best race and not whine for half a year. All that whine could of been hours mastering the new best race.

Tell me the data about people switching to any race during the peak of that race winning at the top level.

Why doesn’t Maru or Hero Marine switch to Zerg now or Protoss when Terrans whined about Protoss for months. How many people switched from Protoss to Zerg on the ladder after they received all the nerfs with more incoming.

Professional players should want to make the most money. Why lose out on money you could make winning with Op Zerg?

None of us can directly make balance changes we think is fair. It would be far easier to just play the OP race then to have endless discussion about it trying to prove it.

Why play the race you think sucks and gives you a handicap. That’s just very stupid.

1 Like
13

So if 99% of the players in a tournament were Zerg, but they only won 96% of the time, that would be grounds for being UP? Just making sure I’m getting this right. 43%? That’s such an insane representation.

As for Terran being over powered vs Protoss, every time Toss wins a game vs Terran it should count as 2 games.

5 Likes
14

ive never talked about people switching races. i talked about strategies and units. why f.e. a serral or dark wouldnt abuse infestors, if they were THAT op.

switching races is a completely different story, since it needs effort to learn a new race, same effort like learning a whole new game. but tbh, for the typical forum whiner, it would widen their horizont like infinitely, if they would try out the race, which is op in their opinion. a pro cant switch races, if they want to stay competitve. since most pros are dependent on their results in tournaments, thats not a realistic thing to do. the normal players on the other side, theres nothing holding you back from switching races. sure you will not be on the same skill lvl at first, but since most of us dont earn money with playing sc2, it doesnt really matter, or hurts just the own ego.

btw my theory to the race distribution: its always said, that zerg is so ez, that you dont have that many zergs at low lvl and so many terrans in bronze.

but what, imo, these dudes dont understand is, that a new player will most likely play terran as first race. why? because it looks the most familiar to other rts games. its not hard to understand for a newbee. want units? build barracks. want mechanical units? build a factory. want flyers? build a starport. from a newbee perspective, thats so damn logical, while zerg and toss is different from what you see in other rts games.

in addition to that - the first campaign is the terran one, so someone who tries out the campaign, will choose terran most likely. if they dont like the competive atmosphere in the ladder, they soon will never play multiplayer again, with letting that poor terran ladder acc to rest.

15

I myself don’t like the over exaggeration when it comes to the balance discussion. You have 3 different races. You have a shifting meta game. You occasionally have nerfs and buffs. You have players with different skill levels who play the game.

I’m one of the last people that will say the game can ever be balanced. Unless you want to remove two races from the game. I won’t deny that units can change in power level when changes are made to the game.

I want the best players to win. If a unit/s is being abused then I would want adjustments to made to keep the playing field as level as possible. I wouldn’t want cheap wins or unfair losses. I’m a Zerg player, if Infestors are to powerful then I hope Blizzard makes the adjustment to keep the playing field level.

I just think people do to many mental gymnastics and overreact to prove what is already known. The game continues to shift. Certain Player/Race/Unit will shine. It will change and new ones will shine. When people fall out of favor then the tears flow heavy.

I’m not directing the accusations of over reaction/exaggeration on you. It does however exist within the community to the point GM streamers are jumping in on it.

No race will carry me to a level I couldn’t do with another race with the same amount of effort. Will there be moment of strength and weakness within my chosen race? Absolutely.

It looks more to me like politics and not actual balance discussion. Until the next election I will sit through “my guys the best” and “It should of been my guy your guy sucks”

3 Likes
16

Hey, Batz! Maybe you missed this question I posed to you on the other thread?

1 Like
17

Except its not a conclusion, it’s something you believe without evidence. While theres a ton of evidence that humans tend to prefer better things, you’re forgetting this isn’t a question of what method grows better crops or how to build strong buildings. People have loyalties and sentiments, which is why people dont travel the world hopping from country to country or people who make huge sacrifices for causes they care about. Human behavior is MUCH more complicated than picking whatever option seems objectively best, and bandwagoning on that. You have no evidence to say that the reason zerg has higher populations is because zerg is op and everyone is switching to it. You just belive that to be true. If you really have a phd in robotics I’m sure you can realize your premises do not lead to your conclusion.

2 Likes
18

You can. It’s called z-testing and logistic regression.

https://en.wikipedia.org/wiki/Z-test
https://en.m.wikipedia.org/wiki/Logistic_regression
https://en.wikipedia.org/wiki/Elo_rating_system

It’s right 99.99999% of the time on average. That means it’s a VERY good way to measure skill.

Yes. Over 500 of them.

ZvZ is going to be the most common matchup when zerg is the most common race played.

Not only Zerg is the most populated in the higher leagues and pro level, even thinking is the least played race in general in the game. Not only, furthemore, the population distribution depicts how Z raise their population with each higher league

This effect exists on the ladder, but not at the pro level. This was thoroughly disproven at the pro level. Furthermore, what you have is a correlation and not causation. If you say zerg is OP because more high-level zergs play Zerg, I could say Zerg is UP because nobody but high level players play it.

The problem with this analysis is that (how SouthWinD said) is that your assumption of “each player has the same probability to win the tournament” is false.
It most certainly is not false. Read:

In probability theory, the central limit theorem (CLT) establishes that, in some situations, when independent random variables are added, their properly normalized sum tends toward a normal distribution (informally a "bell curve") even if the original variables themselves are not normally distributed.

What this says is that with an adequate sample size, anomalies average out. So, if you get one tournament with loads of good zergs, then with a large enough sample you will gets another tournament with loads of good terrans. On average, skill representation will be equal. If skill representation is equal, then each player should have a 50% win-rate on average and a 3.125% chance to win the tournament on average.

It’s a good observation and it’s one of the best arguments for why race selection and skill are not related. People select the race they play YEARS before their full skill is realized. Race switching is very rare.

This is for all weekly/monthly/minor/major/premier events. Literally every tournament in Aligulac. Protoss have the lowest skill ceiling quite clearly.

1 Like
19

I don’t think this way of data usage is what we need. We need every race to have edges and a-symetric advantages. We need to have an harrassment face. A clear defend face. A wow now i have a good army face. Etc etc. Now we just have. Defend max out done. It’s sad how we know the outcome of most games 10-15 mins into the game. You can see it with commentators aswell. Bored af most of the time :frowning: this game needs continious addition of new stuff. So people can’t figure out if balance is an issue. But focus on learning the new stuff.

20

There is no harder/greater skill in this game than figuring out a way to win games with Toss. Most mortals have to give up and just resort to cannon rushing.