Author: Keith Hyams
Date: 04:07:27 02/16/06
Go up one level in this thread
On February 16, 2006 at 00:20:19, Joseph Ciarrochi wrote:
>hi folks,
>
>I am running a 40/4 tournment focusing on spike 1.1. Both rybka 13 b (21.5-2.5)
>and fritz 9 (20 - 4) are killing spike.
>
>I am wondering. At what point is an engines play so weak against another engine
>that it is no longer useful to use that engine for testing. Rybka and fritz are
>winning 89.5%% and 83% respetively. Even if the next rybka was much stronger,
>would this show up against spike, which the old rybka is already killing?
>
>
>best
>Joseph
Hi Joseph,
In a nutshell :-If you want to gain information about the strength of engine A
by playing 100 games against engine B of known strength, 100 draws gives you
more reliable information about the strength of engine A than 100 losses.
Therefore to find useful information, match engines of roughly equal strength.
There is a lot more to say about this topic but as neither of us is a
mathematician best keep it simple.
Remi Coulom, author of engine "The Crazy Bishop" is a mathematician who is
interested in this and he has published work about this on The Internet.
Keith
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.