Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Winboard and Arena GUI test !

Author: Russell Reagan

Date: 15:36:56 01/07/05

Go up one level in this thread


On January 07, 2005 at 13:07:03, Ingo Bauer wrote:

>Hello
>
>In case you are still amazed why the engines do different moves in differnt GUI
>do the following test:
>
>Two engines with a fixed balanced opening position (no books). Now start a 20
>Game match in the same GUI (choose one) with the same position and the same
>color for each engine, of course learning should be switched off and the time
>for each game should not be too short (eg. 15 min each side)
>
>You will see that you do not get 20 identical games. Not all moves are allways
>reproducable, most chess engines are not deterministic and it is gettiing worse
>if you use dual engines.
>
>Bye
>Ingo

If there are nondeterministic single-cpu chess engines, then maybe a better test
for GUIs would be to use simple, open source engines. Do you think a simple,
open source engine such as TSCP is more likely to be deterministic?

Or does the issue of time cause undeterministic behavior no matter what? i.e.
maybe a background process (system process, firewall, antivirus, etc.) randomly
takes enough CPU time away from a chess engine to cause nondeterministic
behavior?



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.