Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Dynamic Pruning System

Author: Jonathan Lee

Date: 20:33:29 12/20/02

Go up one level in this thread


On December 20, 2002 at 23:17:29, John Lowe wrote:

>On December 20, 2002 at 11:49:17, Jonathan Lee wrote:
>
>>Statistics are interesting computer chess.  For example, you
>>can prune the nodes differently (one for the middle game, another for the
>>endgame, another for the king and pawns only, etc.) so that you could have a
>>dynamic pruning system.  (Does that make sense to anyone?)
>>
>>Or do you think efficiency has almost reached it peak in the software end of it?
>>Jonathan
>
>I've noticed that some problems are solved faster with shallower searches. - for
>example if you have a lone king against a king bishop & knight the machine is
>best guided by heuristics rather than letting it evaluate thousands of
>anti-heuristic positions. Is this the kind of thing you're talking about?
Mephisto Genius 2 can solve the K vs. KBN.
Genius 2 can be found at http://www.gambitchess.com/progr.htm    the one with
114 kilobytes (or 54,313 bytes the executable).
Genius 2 does not use null move, therefore pruning isn't a big problem here.
Jonathan



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.