Originally posted by no1marauder
The proper procedure has been followed regarding this user years ago. The evidence then was overwhelming and every analysis done since has duplicated the original evidence and results in dozens of games.
This is all down to "politics" now.
Every analysis? What analysis? This stuff isn't published - at least not here.
I am not interested in hearing that you analysed x games and found a match up rate of y, This means zip. You have to demonstrate that the probability of a player of a similar standard would get the same match up rate is lower than some threshold value. Since the number of good moves in a position is on average pretty low (probably 2 or 3 depending - an assumption has to be that they are equally probable) you need a very high match up rate to hit a p-value (probability of getting the same result by chance) of 0.01or 0.05 or whatever you decide is the threshold for beyond reasonable doubt.
You also have to take into account that for the top 5 or so on the site it is rather difficult to get a control group to compare their match up rates with. There are also a number of other potential sources of bias you have to control for. An obvious one is the possibility of opening theory which the games mods are unaware of - basically you have to hope that the averaging procedure corrects for that. In the less obvious cases you would ideally get a player who is sufficiently strong to establish candidate moves (and show their preference) in a subset of games to ensure that your estimates of moves available per position are correct.
I assume (hope) that the games mods do their analysis using this kind of statistical rigor. Which of course takes time. Copious amounts of it; another poster in this thread, who had done some independent analysis of someone's games which they discussed with me using the pm system, told me that it takes 1.5 hours to get the basic statistics on a single game - never mind the subsequent stuff I'm talking about.
I strongly dispute your case that the "analysis has been done", at least with sufficient rigor for the site admins to proceed. Even if it has this is not the point. Throwing allegations around about specific players in the forums really does not help anyone. I seriously hope that Russ will not act against players because a few people want him to and I also hope that he will not fail to act when no one is pressuring him to.
I do not care about the specific outcome of this. My point in this post is that you have to be careful about statistical analysis, hacking around with a couple of dozen games is not sufficient. My general point is that it is
against the rules of the site to accuse specific players of cheating in the public forums for reasons that have been discussed ad nauseam in other threads. The rights and wrongs of this particular case have nothing to do with that.