Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Correction to Ed's page about null move

Author: Uri Blass

Date: 14:33:37 01/06/03

Go up one level in this thread


On January 06, 2003 at 15:54:48, Igor V. Korshunov wrote:

>On January 06, 2003 at 03:48:22, Uri Blass wrote:
>
>>Ed writes in his page in selective search techniques:
>>
>>"Nowadays I can't mention a chess program that doesn't use Null-Move, the chess
>>programmer community owes Frans Morsch a big thanks."
>>
>>Junior does not use null move so it is possible to mention chess programs that
>>do not use null move.
>>
>>Uri
>
>About two years ago I red that Junior can make mistake in endgame positions due
>to very heavy null move pruning.


Junior(at least the public version) can make stupid mistakes in some positions
but the reason is different.

The reason is wrong knowledge about draw by insufficient material or wrong
knowledge of stalemate detection or some wrong assumption about repetition
detection(Junior seems to detect possible draws only after 3 plies of the form
from1->to1,from2->to2,to1->from1 without checking that to2->from2 is legal)

Usually this trick help it to be faster but there are rare cases when
to2->from 2 is simply illegal.

An example is Rf7-f6 Qe6-g8 Rf6-f7 and the queen cannot go back to e6.

I guess that Junior7 assumes at least draw for the side with the queen when it
analyze this line.

My guess may be wrong and maybe I am too fast to generalize from one example
when junior blundered and showed 0.00 score but this is my guess.

I posted some positions when Junior blundered in the past and if people are
interested they may search for them in the CCC archives.

I do not have them in this computer.

Uri



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.