It’s not my data. It’s from teamliquid.
Yes, but is this data from 2023-2024 or 2010-2024?
Given protoss performance trends in aligulac and the sc2 grandmaster league, we’d expect protoss overperformance to extend back to 2018. The first tournament included in this set is from 2020, so well within the confines of protoss overperformance trends as recorded elsewhere.
We have GM representation, pro level win-rates, skill metric skewing, race switching statistics, etc, it’s beyond reasonable doubt that protoss is favored in the current game design. The uniformity of many different statistics meets the specificity requirement of the bradford hill criteria a second time over.
There were indeed some tournaments, where protoss had 7/8 Play-off players back then, but it seems there were quiet a few nerfs to protoss, which changed that, since then.
Protoss performance went down very slightly from where it was at one point. They set record high winrates in the pro scene not too long ago (60.44% PvZ) and record high GM numbers (45% of GM) but now are only 40% of GM. Their performance is still far too high to explain with a balanced-game scenario. We’re talking one in a trillion trillion trillion odds that the game is balanced.
The game is not balanced. Every time a pro-gamer finds a new build, it literally trashes the balance and forces everybody to search the solution or call for balance patch if they can’t.
Yes, protoss had very high winrates back then. But the balance patches did it’s job. There are not that many korean protoss players who are able to get into play-offs, while they are multiple european terran players who are getting there just fine, which is a total reverse of what was back there.
There have been very few changes to the build orders / strategy of SC2 in at least 7 years. Serral won IEM by using 2 prong hydra bane which is what Leenock made a deep run in the GSL with 7 years ago.
Do you have eyes? Please read the various statistics before posting clearly false assertions. They are 40% of GM – almost 2x as numurous as zerg, for example. Protoss have high winrates right now. The balance patch did not do its job.
Do you mean there are 80 protoss, 60 zergs, 60 terrain players in GM? That is a huge difference to me. I mean whoah, 20 players. That is hell as much.
Maybe leenock did something 7 years ago. But when I see how Serral is constantly dropping queens on the map to speed up creep spread, I can’t remember those tricks 7 years ago.
So I wasn’t talking about huge changes. Even something small like a timing due to a patch which changed the price by 25 minerals can make a lot of difference in pro scene.
A) this is broader than gm league.
B) u r being sarcastic but it matters to the zerg gms who are denied the accomplishment.
C) there are 3x gm leagues, so u underestimate the number by a factor of 3.
D ) balance issues make ZvP needlessly difficult in every game played.
a) Do you mean it broads into masters? Well, yeah, then you have to talk to terrans, who are the 1st on every server.
b) They are not really denied anything, they are nerfed, because of Serral. And they have the least number of players on the ladder on top.
c) Yeah, I totally forgot that is actually 60.
d) Do you mean on ladder or in pro games? Because we know, that ladder is rarely getting balance fixes.
Argument to irrelevance is self defeating because it devolves into a special pleading argument. If it’s fine to deny zergs a gm promotion then it’s fine to deny protoss and there wouldn’t be an issue in buffing zerg to equal gm representation. Why the special treatment for protoss in your reasoning-- I think we all know why.
pic unrelated
https://i.imgur.com/vdkPKms.png
Why can’t you read the whole sentence for once? I literally said, that zerg is being nerfed not because of PPP whine, but because of Serral. A factor which ladder players like me, have no influence on. So I and other protoss players can’t do anything to help zerg on the ladder. It is not being denied by us.
Every race is easy and hard, depending on the map, match up, your level of play and tons of other mitigating factors.
1 Like
I played all 3 races and I think zerg is the least demanding. The fact that you can make the whole production cycle without switching your attention to it makes it great to focus more on every task. Also this race is super forgiving, I remember a game when I lost my whole army in an awful engagement, pressed 1 button for 3 seconds and counter-attacked with success. With protoss or terran my production would get overwhelmed before I would gather the new army.
That depends on whether it is a universal trait for the entire faction, or a trait for some particular units that may or may not apply only when a certain amount of micro (or some other skill) is involved on either side of the match. In the latter case, there definitely can be a yo-yo effect where different races have the advantage at different skill levels.
Map-design is also a factor in this, since the performance of different unit types is often influenced by map features, and not all of a faction’s units will be affected the same way.
2 Likes
That’s why the specificity requirement exists in the Bradford-Hill criteria. IF skill metrics are race specific, or level specific, we should see protoss lead in one or two, zerg should lead in one or two, and terran should lead in one or two. Unfortunately, protoss lead in none. Clearly this is specifically a protoss issue, hence the specificity requirement is met.
Moving your screen, inputting actions, keeping your bank low, etc, are universal skills that apply to all three races, and it’s absurd to say otherwise. Protoss is simply easier to play.
The problem with this theory is that we can measure, objectively, how hard it is depending on these various mitigating factors. The only common factor is protoss, ergo ease of play does not correlate with any other factor other than toss. Translation, your theory is opposite of the data. If ease of play correlated with league, and not eztoss, we would expect protoss to have identical skill metrics for all levels of play on the ladder. This would show invariance to race and variance to league. That would prove your theory, but that’s not what the data says. The data says all toss, regardless of league, have lower skill metrics for given performance level. This shows invariance of performance level and variance with race, ergo race causes performance. The only thing that causes this is if the protoss group gets a performance boost that upsets the skill metric correlations. Toss gets you a little bit higher on the ladder than the same skill level would achieve with zerg or terran. That’s how you end up with protoss losing 4 oracles, in a single game, in major tournaments. They just don’t have the same level of multitasking skills.
No, you shouldn’t.
Do you seriously think Reynor’s mechanics drop by ~20% when he picks Protoss compared to when he picks Zerg? Do you think he just become 20% worse as a player when playing Protoss?
Differences of 20-50% APM playing different races/styles from the same player, depending on skill level, are normal.
Those skill metrics measure differently depending on how the player is playing, what units they’re making, etc… and how their race works affects that, which is always the problem with comparing APM between races.
There’s also another problem that you’re ignoring: Things aren’t the same at all levels. Diamond is flooded with Zergs, for example (it’s the most represented race in Diamond, despite the lowest overall representation), while having extremely low representations (and high win rates) in silver/bronze.
Starcraft ladder stats: https:// sc2pulse nephest com/sc2/
In GM and Masters, Zerg, despite being the least represented of the 3 races, is actually still represented above their overall representation as a percentage. In fact, in GM, the relative representation with respect to total representation is practically the same for both Zerg and Protoss right now.
That indicates that there is likely some breakpoint around that diamond skill level, where things get relatively more difficult for Zerg compared to the other races, but something about it makes it easier at the lower leagues.
Protoss has pretty flat representation in all leagues, all being very close to its overall representation; a bit higher in GM and gold league, and a bit lower in diamond and bronze (which is nearly 50% Terran).
A middle league player finding Zerg to be the easiest race shouldn’t be a surprise given those ladder metrics when looked at more closely. It’s actually an expected outcome. Whether they find Protoss or Terran more difficult at that level likely depends mostly on their skillset.
3 Likes
By your “logic”, a balanced game state would have practically the entirety of GM as Zerg, and only a few of the very best players would ever be GM as Protoss, because people playing Zerg have the most APM (since you completely ignore that the APM of Zerg players drops dramatically if/when they decide to play Protoss or mech Terran instead, and even bio Terran still has notably less APM).
2 Likes
That’s your logic, not mine, and it’s pretty inane to be frank. I’ve said repeatedly that the skill metrics should be randomly distributed between the races, meaning it’s fine if zerg leads in apm and epm if toss leads in something like hotkeys or screen movements. So using apm as a proxy to predict future gm performance is clearly a deliberate straw man of my argument. If anything I’ve said, very specifically, that APM should not predict racial performance in GM. You just need to actually read what I write instead of superimposing your bias because, frankly, this is absolutely ridiculous.
What on Earth makes you think that in the first place, because it’s patently insane. Reynor’s mechanics go down, when he plays eztoss, because eztoss requires less. Pro level games are insanely fast and so having less stuff to do is a huge advantage.
You’ve never looked at the relevant data in your life because to even acquire the data you’d have to be a data scientist. So no, you have no clue if 20-50% apm blah blah is normal.
Blah blah nonsense.
My theory properly describes everything from bronze to serral. You can try to make counter points and I will try to explain how these systems work but honestly this is going to be a fruitless effort.
Players race switch as a result of balance and that’s a proven fact. It’s not disputable. If you calculate the correlation between win-rate and race switching you’d see that zerg’s numbers have down on the ladder proportionally to how overpowered protoss is. Basically there is a strong correlation between protoss in gm and the number of zergs on the ladder, it meets the bradford hill criteria, and so it’s a causal relationship. This isn’t up for debate. Low level zergs either quit the game or race switched because they were getting absolute wrecked by eztoss.
When XvZ favors X, X moves up the ladder, Z moves down the ladder, and the lowest level Z players race switch or quit. The result is tighter variance in both X and Z populations (in X the variance goes up over time due to race switching), higher mean performance in the Z population, and lower mean performance in the X population. You heard that right: the overpowered race has lower mean performance (due to race switching). More people play X, and fewer will play Z. All of these factors are measurable and present in the current population. We have higher PvZ winrates. PvZ favors toss. Toss are higher on the ladder than their skill metrics allow for, zergs are lower. Zerg is the least played, toss the most. The variance is tighter for zerg than it is for toss. Zergs mean is higher than protoss’ mean. Etc.
The sc2 ladder is a point distribution that follows rules and we change the distribution by changing the rules. The rules that are changed is how favored toss is vs zerg and how dependent the probability of race switching is on how favored a race is. You change those factors and you perfectly describe the sc2 ladder all the way down to bronze and all the way up to serral.
Blah blah nonsense.
Unfortunately for you “flatness” isn’t a meaningful statistical measure.
Sorry kiddo, subjective measures do not outweigh hard data. I can’t believe I have to say this, but we live in 2024 where we know there is an objective reality that can be measured and quantified. Your feelings do not outweigh the facts and your beliefs do not trump the science. The data says what the data says, it doesn’t say what it doesn’t say, and that’s that.