Author: blass uri
Date: 04:33:08 07/29/00
Go up one level in this thread
On July 29, 2000 at 06:24:51, Steve Maughan wrote: >Terry, > >You can download a small utility (WhoIsBetter) from my web site that show the >number of games that are needed to be statistically significant win > >http://www.maughan.clara.net/computerchess.htm > >It's amazing how many are needed to be 95% certain of superiority! > >If you want to dig further, the statistics used in the program are based upon >the binomial distribution. > >Regards > >Steve I think that a better program should also give the user to insert probability for both sides to win. Your program assumes that white has probability of 50% to win and the same for black. Your program does not answer what is significant result assuming probability for white to win is 40% and probability for black to win is 30%. Uri
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.