Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Statistics and Test results

Author: Chris Welty

Date: 08:38:41 10/07/04

Go up one level in this thread


>However, since the sample isn't random, the entire test is meaningless.

What makes you think the sample isn't random?

>If you could test engines like that, you would use the binomial
>distribution and would need more than 30 random games from those engines to
>properly test the probability of one engine winning over another.

That's just wrong. The number of games you need is dependent on the results of
the games.

> However,
>since it is not really possible to get a "random game", you will need to play,
>as others on this board have suggested, you will need to increase the sample
>size to 1000 or so.

Again, that's wrong. The sample size you need depends on the outcome of the
games. If after 500 games the result is 500-0 would you agree that one engine is
better than another?



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.