Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: To R.Hyatt re Crafty,s playing strength

Author: Matthew Hull

Date: 09:09:24 07/12/04

Go up one level in this thread


On July 12, 2004 at 11:35:05, Kurt Widmann wrote:

>Congratulations for crafty's improvements.
>I am not a programmer,just a cheerleader for the under dog crafty
>and always hoping that your program will be an equal to Fritz and Jumior.
>In this tournament I noticed a glaring weakness of crafty as to the
exploitation >of its own strength ie protected passer and also the reluctance
to acceptance a >draw against same or higher rated opponent which resulted in
the loss of 1.5 >points.
>It seems to me that crafty needs instructions to remove these weaknesses.
>In two games crafty was missled to attack an appearant opponent weakness
>instead of playing to its strenght,in both cases it was the c-pawn.
>Why not try to instruct your program to persue searh in favor to advance
>its own protected pawn if it can not find a value gain from going after an
>opponents weakness within and or go into a defencive mode at a certain horizon
>limit. And also take into consideration the rating of its opponent in an
>appearant equal position to secure a draw first before imparking into
>uncertainties.
>Keep reaching for the top, Kurt Widmann a former ICCF player.


Not too bad a showing for having no great hardware advantage vis-a-vis the
commercials.  Massive scalability it's  principle equalizing feature.  It
did not get a chance to show it in this contest, but still finished
respectably.

On the other hand, the ParSOS game was pretty dismal.  Crafty either has a
search bug or it needs a major eval tweak.  It wandered into some rather ugly
positions in this event, which I've noticed is something of a bad habit on ICC.



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.