Computer Chess Club Archives


Search

Terms

Messages

Subject: 6 draws say better about a machine than 3 losses and 3 wins

Author: Rémi Coulom

Date: 04:10:59 02/17/05

Go up one level in this thread


On February 17, 2005 at 05:03:17, stuart taylor wrote:

>There are a few of them on this and recent pages, and some ideas are similar to
>things I've written earlier. But what _I_ consider to be my best posts are
>usually simply ignored (seemingly). I hope that means that people agree with
>some of my points. (like that 6 draws say better about a machine than 3 losses
>etc. [because losses necesitate weakness, whereas draws, maybe yes, maybe not]).
>S.Taylor

Hi Stuart,

I have not noticed your post about 6 draws. You might be interested to know
that, in the Bayesian inference experiments that I am currently running, I use
the statistical model of Rao and Kupper, in which one draw is equivalent to one
win and one loss. My experiments indicate that this model fits the data very
well in the case of computer games. So, it seems that statistics confirm your
intuition.

Also, regarding the number of replies, it is usual that the most interesting
posts do not get many replies. Please don't stop posting whatever you find
interesting.

Rémi



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.