But then (by your logic not mine) I can claim any variable accounts for all of the skill because they are not possible to differentiate. It does not go all the way but it is enough to show that your claim is rush, and that statistical tools cannot account for the whole poicture
1 Like
Correct. In that case APM is not differentiable. So youâre using another word for the same variable. So what you are doing is renaming the variables then saying âSee itâs not APM!â.
This in logic is called a tautology, which are not really admissable in an argument. When you say everything, you say nothing. You are just saying skill with another word, and that emotional state is also skill 100%
You are describing tnothing.
1 Like
I donât think you really understand the implications of saying APM = Skill. If that were true, there could be NO exceptions; that claim is an absolute. That player in plat with 200 APM and a <50% win rate? They donât belong in masters.
Currently, youâre almost certainly right that Protoss in GM, are, on average, less skilled than their counterparts. I donât think anyoneâs even tried to disagree with that claim, just your method of âprovingâ it. That has little to do with APM; just looking at the number of players of each race in GM, and their average positions in it, is enough to make that claim.
Thanks for admitting you are intellectually bankrupt and have no way to undo my mathematical proof. That is why it is called a proof, after all - it proves something.
Youâre making things more complex than they have to be. I think protoss economy and units are just too strong. APM may have some correlations with skill and with some of the things youâre mentioning, but its much simpler than that. Protoss is just too strong.
My claim is that you have no mathematical proof whatsoever, because your method cannot account for skill, since it cannot even tell apart the variables.
Kid you said youâve taken Stats with Calc so if you canât understand that a variable conjoined with APM cannot have a correlation with MMR given corr(APM, MMR) = 1, then youâre admitting youâve never taken calc nor stats.
It is mathematically impossible for a variable to be conjoint, independent, and a have correlation of 0. If corr(APM, MMR) < 1 then you could make the argument that other variables exist however itâs 0.998 meaning the maximum a variable could correlate with MMR is 500 times less than APM. In other words, it is proven that APM=MMR.
1 Like
What you do not understand, is that you dont have a math problem, you have a methodology issue. I could have a 1 to 1 Corr between between the number of cats and the number of fish in the sea if I pick the right samples.
2 Likes
You are violating the conjoined requirement in your example. This isnât rocket science, people.
The issue with your mathematical models is they never account for enough variables. As others mentioned, EPM, races simply needing more APM or differing APM movements, what the APM is used for, such as creep spread vs. micro, and so on.
Wrong. I just proved it is mathematically impossible for any other variable to exist that is both differentiable from APM and that has an impact on MMR which means I have accounted for 499 out of 500 parts that make up skill.
Ok I give up for real now, I just hope you are not planning as a career to be part of any sort of investigation that affects public policy, then we would really be in trouble.
1 Like
Sorry kid but not understanding what âconjoinedâ means is not a valid argument. Everyone agrees that APM affects skill. Nobody would deny that. The question is how much. So we have a variable that we know is casual with skill and correlated. In your example you list a variable that only correlates but that is not casual. Thus you violated the conjoined requirement.
Just to break something down here:
You canât use averages when making a claim that Skill and APM are equal.
They are strongly correlated, yes, but using averages will strengthen the apparent correlation by ignoring the outliers that, by their very existence, make your claim wrongâŚ
The claim also ignores that there are differences in the races that cause significant differences in APM. That person with 77 APM in gold when playing Protoss will probably have 110 APM when they play Zerg.
Do you seriously think that that average gold league Protoss is ~70% as skilled as the average gold league Zerg?
1 Like
He could do it on NA, not long ago I saw Golden winning a game with no keyboard (left hand on his head) against a 5,5k player.
LUL⌠op have slaughtered arguments. Before troll have posted themâŚ
Strengthen of correlation is proof. Of op point⌠troll say can not do it? LULâŚ
1 Like
Pro players have 300 APM and EPM is 200 or above. If you tell me that this means nothing, go home because no noob on ladder does that. Even when they claim that there are lwoer ones with 200 APM. When they say APM matters they mean how it looks also with EPM
1 Like
So just to clarify, you and bobayoga believe serral can beat sub 6.5k players like puck and scarlett with 100 apm?
3 Likes
My point is that it isnât proof of the claim APM = Skill.
It only proves a fairly strong correlation, though his use of averages rather than individual data points, made that correlation look even stronger, and nobody disputed the idea that thereâs a strong correlation between APM and skill in the first place, so thereâs no reason to even try to make a proof of that.
His claim of APM being Equal to skill is an absolute one - if there are ANY outliers, then we know that itâs wrong. Using averages means we canât see outliers.