Author: Roger
Date: 20:42:07 01/27/00
Go up one level in this thread
I find your post to be very interesting, because you are essentially saying that a certain level of variability is intrinsic to the results, and that even the loser can't be judged worse than the winner unless the difference in their scores exceeds some threshold. I guess that's why the results of so many tournaments are often so different, and why the results of any one tournament are so suspect. For whatever it's worth, the binomial theorem gives the probability of a particular result where only two outcomes are concerned. You could probably compute the probabilities of a particular result with draw included using a chi square. You'd need a better estimate of the frequency of draws, of course, in order to obtain the "expected frequencies." You could then plug in whatever "obtained frequencies" you wanted, do the math, and then look up the probability that such a result has occurred by chance alone. Roger.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.