APM is an accurate way to measure skill because it has a strong correlation with skill. Your story that “APM=skill” was confusing doesn’t work when it’s been repeatedly clarified in the thread. You keep saying the opposite of what has been clarified. You’re trying to force the conclusion that I am wrong when in reality you simply were confused.
APM may very well be near equivalent to skill however my point is that my argument DOES NOT rely on that being true. My argument only relies on there being a correlation strong enough to make APM a good predictor for skill, and that’s it. It’s been clarified too many times at this point so saying otherwise is just a lie.
Let him tell us that (considering that he liked your apologetic manure). He is on record of saying otherwise.
On the other hand, the implications and conclusions that the clown has drown from the apparent APM disparity point 100% toward causation.
You misunderstand the very nature of the claim. “APM correlates with skill which makes it a predictor for skill” does not rely on causation to be true. Causation is not even relevant to the claim.
We could have a conversation about the true cause of Protoss’ lower skill, but that’s not what this thread is about. This thread is purely about proving that Protoss does, in fact, have lower skill. It does not prove the cause of that lower skill.
…and that’s means nothing because APM disparities are due to differences in race-design. It’s expected to be different. People have made a long list of things that constitute skill where APM is a minuscule thing compared with game-knowledge, judgement and other factors. You insist that APM is alpha and omega of skill in SC2 …because everything is reduced to …actions.
Your “argument” was obliterated from CollegeWings and you as a sore-loser redouble each post.
Pathetic.
One has to understand when time is to let it go.
When I play protoss, my apm is naturally much slower than when I play terran or zerg because that’s simply how the race operates, it requires less APM.
I just think Protoss economy and units are too strong. That simple.
Saying protoss having less skill because of less apm is such a batzy thing to do. Its pretty sad that in 2021 zergs actually still rub one out because its the race with the most apm.
If we only look at correlation between apm and skill within 1 race the correlation will be much higher. Because comparing 3 different races with 3 different mechanics and playstyles is completly nuts. We also could compare the apm within one race with certain playstyles. Like mech Terran vs bio Terran for example.
When i Switch to zerg and get my 50 free apm i always feel like serral <3
There’s literally no sense in what you’re saying. Either he found a coincidental relationship between two variables and cannot draw any conclusions like you’ve said (thus debunking his conclusion that apm = skill, making him a complete fool) or you have absolutely no idea what you are talking about and this OP clown has yet to explain himself on literally every single red flag that has popped up on his analysis. I literally cannot break down this explanation any further than this. I can’t help you if you are too stupid to understand simple English. Either case makes OP look like a fool, but it’s clear the second one holds true as you’ve made it more clear you have no idea of what you are talking about. I’m not going to entertain your delusions further if you are just going to blindly say “sorry but no” when it’s clearly a yes situation. What you are doing is the equivalent of walking into a room and saying the sky is red when people are just having a random conversation. Just completely unnecessary noise and false information. As a bonus, neither of you have addressed any flaws regarding the statistical analysis OP has attempted. I wonder why?
I’m literally quoting what you’ve said. I’ve asked questions you’ve refused to answer. I have found holes in your logic that you continuously avoid in this thread and the other. I am not forcing the conclusion that you are wrong… I have all the increasing evidence that you have no idea what you are talking about right here and in the other thread.
Okay so let me get this straight, and I’ll try to explain it as simple as possible so your pea brain understands what is wrong with what you are saying right now. You LITERALLY say APM = skill. Then you bring up some numbers and shaky analysis to try and back that up. Now you are saying
when in reality, the only trend you have found is that MMR and APM have a positive correlation, meaning
You cannot draw the conclusion that APM is a good PREDICTOR for skill
The fact that you keep pressing the issue that APM = skill while apparently having 0 solid evidence to back that up makes 0 SENSE. And I know as I say this you are going to try to draw me into an endless loop of stupidity as you try to cover up for your errors…
Everything you said in the other thread has lost its meaning
Never mind the fact that
You refuse to address problems in your data points
You refuse to address the problems in your analysis
You don’t have any explanation for all the outlandish assumptions you are making
Making the correlation you found between MMR and APM (NOT SKILL AND APM) very sketchy. I CAN’T break it down any further than this, and I am not going to repeat myself for you or this other clown that showed up out of nowhere to insert his irrelevant delusions into this discussion.
It is exactly this, which is why I am confused as to why OP is liking Gabriel’s posts. Neither clown has any idea of what the other is talking about, and I’m not going to let either of you drag me down into a spiral of stupidity until OP explains all the flaws in his analysis.
Deeply disagree with OP, but correlation while not 100% reliable (though still very useful as causation implies correlation) for stating causation, it can actually be a very good predictor depending on the type of correlation without the need to prove causation. That is why in statistics certain variables are often called predictors instead of causes.
A good example of this, is the rise of demand in soda and ice cream is a classic example of correlation, the cause in this example is the rise of temperature due to the summer. Are soda and ice cream cause of each other? no. Are they a good predictor for each other? the answer is yes because they have a common cause, now this will change in a multivariable analysis that adds different causes for each other.
Also, the phrase “Correlation does not prove causation” is often overstated to prove more than what it actually is trying to say which is that “correlation alone is not prove of causation”. Because if you have a strong correlation it is appropriate to at least suspect a causal relation.
My take on the matter is: Denying that APM is influential to skill level is asinine. Stating that APM is the chief measure of skill is also asinine.
Self translate action per minute these action can be effective or just empty spam keyboard key questions is how much player just spam witch low efficency and how much do low apm but it more profit then you just spam
The point of this thread is to prove that there is a strong enough correlation between APM and MMR to measure skill. Measuring skill we can prove that many Protoss in GM are Master-level in skill, thus proving Protoss is imbalanced. I do not need to prove causation, e.g. what causes Protoss to be lower skill - I only have to prove that they are, and I have. That threshold has been met and it is now fact.
Now if you want to talk about causation, I will refer you to page 11 of the AlphaStar paper which definitively proves that APM is a scalar for your rank:
https://i.imgur.com/kg8EINO.png
According to this chart, a 50% decrease in your APM nets a 3x reduction in win-rate compared to an equal opponent (equal in all respects save for the APM disadvantage). Your win-rate goes from 1:1 to 1:3. So APM scales and caps your maximum performance. You can have GOD-tier game-understanding, but if your Apm is only Platinum league your MMR will be Platinum league.
Which is common sense as probability theory dictates. A person at an APM-deficit compared to their opponent will be at an increasing likelihood to make a game-losing mistake, and this probability will grow with a larger APM gap.
So there are multiple things that affect your rank but they are multiplicative and not additive. E.g.,
IDK if I am misreading the chart, but what it appears to imply is that APM is important up to the point where you reach 50% win rate, and becomes irrelevant after that. Maybe you could give more context because if that is so, the case you are making simply falls.
Also the chart cuts both ways. Because it is also the case that if you have 500 apm vs 200 apm. But the rest of your skill is not up to par. Then APM is simply irrelevant. So you can lose to a player with lower APM than you with him being more skillfull.
I also wonder how the study (dont havee much time to look at it), accounts for APM differences between the races that clearly do not account for skill E.G rapid firing lings and Corrosive Biles. Because you would have to normalize them all as one before you start comparing them, and that seems pretty hard to do.
Fianally you are assuming APM is the weakest link, and I would argue that APM might better work in your model as a Beta (maybe you are implying that) and even as a base variable. Since it is a minimum requirement for skill and it multiplies the other factors (caps them), but not necessarily the main differentiation once its present.
There is a lot of iffy things, with your conclusions (not the study) which seem rush. Statistics are useless numbers without interpretation.
Because there is a certain amount of APM needed to play the game, and once you reach that limit adding more APM doesn’t help your win-rate.
Please see:
Each race had an APM limit similar to the APM limit of humans who played that race. Given that APM budget, AlphaStar Protoss was vastly stronger with Protoss than with Zerg. That’s a confirmation of these findings in this thread. Zerg is least in GM despite having higher skill metrics (APM etc) and Protoss highest in GM despite having lower skill metrics. That’s how it was for AlphaStar. Their Protoss had lower APM yet higher MMR.
Protoss has a massive design flaw. It is much easier to play and this makes it difficult to balance it across a spectrum of skills.
Yes but my point is that capping them at the same apm is not a good method to determine which race is easier, you would need a differentiated cap, because you simply rapid fire more as zerg which makes your apm higher while not really proving any skill.
Also think of it this way, not all actions are as easy. And that is Qualitative and not quantitative. It is not as simply as assuming the same actions across the races.
