Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Rybka is not perfect! It doesn't see mate in 4!

Author: Richard Pijl

Date: 02:26:21 12/06/05

Go up one level in this thread


On December 05, 2005 at 20:47:17, Joseph Tadeusz wrote:

>On December 05, 2005 at 19:14:46, Martin Giepmans wrote:
>
>>[D] 8/3N4/8/8/1pp5/brpp4/qpprp3/1nkbK3 w - - 3 13
>>
>>I guess many other engines will also not find it. Mine only finds it when I
>>enable verification search (which is default off).
>>It's a null move problem.
>>
>>Martin
>
>I think this position is perfect for double null move.
>
>The baron, which uses it, finds it instantly.

Sorry, but in this case that has another reason, and also shows a requirement of
doublenullmove in finding zugzwang situations.
For zugzwang detection to work with double nullmove, all decision criterias for
doing a nullmove or not should be symmetrical for side to move. Meaning that if
you want to nullmove only when the side to move has at least a piece (which he
can use to 'lose' a move), you also have to require that the opponent has at
least a piece.
I noticed that in knight endings there is often a zugzwang motive that is
decisive in winning or drawing a game. So in the Baron a single knight is not
considered sufficient material to allow nullmoves.
In this position it means that nullmove is turned off completely.
Richard.



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.