Computer Chess Club Archives




Subject: Re: A New Approach to Draw Detection by Move Repetition in Computer Ches

Author: Walter Faxon

Date: 15:07:55 07/30/04

Go up one level in this thread

On July 30, 2004 at 06:47:39, Gerd Isenberg wrote:

>On July 29, 2004 at 23:18:53, Walter Faxon wrote:

 [ snipped ]

>>Will this detect when two like pieces have "traded places" in the repeated
>Good point.
>I don't see how the "New Approach" handles "traded places" as well, because the
>list_of_moves doesn't contain piece information but only from/to squares.
>So occasionally the "New Appoach" may miss some repetitions, where rooks or
>knights have traded places. Whether this is practically relevant is another

And multi-queen endings (like from a Q+P ending).

It's certainly OK if you don't catch all repetitions in quiescence search --
some programs don't check for them at all in QS -- but only there, and only if
you know that you indeed might be missing some.

 </Pedantic off>

-- Walter

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.