Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The Battle of the Crowns has finished!

Author: Franck ZIBI

Date: 08:52:43 01/12/01

Go up one level in this thread



Hello Mike,

When the tournament started, the latest official ZChess release was ZChess 1.2
from ... february 1999 .
It was this 'young' ZChess version that played in the tournament.

At that time I was still thinking that a 'Lot of Chess Knowledge'
was better then search speed.
This ZChess 1.2 was computing thousand of chess specific things ,
like the kind of opponent pieces attacking each square of the board,
the kind of squares (attacked, occupied, defended, ...)
in front of each passed pawn, etc...

I eventually realized that, for ZChess (I do not know for others)
this huge eval function was adding a lot of 'noise' to the program evaluation
(somethink like polluting the eval).

So from version 1.4 and +, I slowly became an adept of the KISS principle
(Keep It Simple Stupid).
This had some strong positive effects on ZChess:
Version 2.22 is around 150/200 elo point stronger then ZChess 1.2

Best regards.


>Referring to Yace and Zchess: After what I've read and heard recently about
>these two, it may be quite unlikely that they would be and the end of this field
>in their most recent versions IMO - but on the other hand, most of the others
>will have been improved too meanwhile...
>
>Maybe somebody finds remarkable positions within their games, where the new
>versions play better. Each such case found would provide explanation for the
>somewhat surprising defeat, and furthermore show where improvement has taken
>place.
>
>Regards,
>M.Scheidl



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.