Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Rating via test positions

Author: Dann Corbit

Date: 12:22:55 10/04/99

Go up one level in this thread


On October 04, 1999 at 14:12:24, Sven Reichard wrote:

>Last week somebody asked about testing programs/hardware using sets of positions
>instead of c/c matches. It is correct that Bratko-Kopec contains only 24
>positions, but a huge collection of problems is in the distribution of Gnu Chess
>v.4 (they've not included it into v5), available at
>
><a href="http://www.fsf.org">www.fsf.org</a>.
>
>If people are interested, I would volunteer to coordinate testing some
>soft/hardware combinations and publish the results at my homepage. Unfortunately
>I don't currently have any powerful computer, but this will change within a
>month or so :).
We have already analyzed all of them (and many more besides) at 12 minutes of
PII 300MHz time equivalent.  Altogether, about 8000 EPD positions were analyzed.
See:
ftp://38.168.214.175/pub/Public_CAP_Results/Apocalypse/



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.