Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: DB doesn't do NULL move????

Author: Robert Hyatt

Date: 18:24:10 06/28/98

Go up one level in this thread


On June 28, 1998 at 20:32:58, jonathan Baxter wrote:

>So all this recent discussion about DB says it doesn't null move, and I saw
>another post which said in some positions it was getting no more than 11-ply.
>Since PC based programs now reach that easily (without heavy forward pruning
>apart from Null move), but PC programs are not close to challenging Kasparov,
>can't we only conclude one of two things:
>
>1) DB has a much better evaluation function than the PC programs.
>
>and/or
>
>2) Null moving makes *lots* of mistakes.


without a doubt.  It lets you find some things, but it screws up badly
on others.  That's the risk of "forward pruning" of any kind...


>
>
>"1" seems unlikely in light of a quote I heard from one member of the DB team to
>the effect that DB's eval was only "nearly as good" as the best commercial
>programs.


I believe this was an *old* quote.  They have since said, on several
different occasions, that they believe they are far better (evaluation-
wise) now.  This was true with the first deep thought/deep blue chip
I think, but the new chip used in the last match had a truly significant
increase in chess knowledge.
>
>So that leaves 2. This is news to me. Anyone have any comments?


null-move is speculative.  Cray Blitz is tactically stronger than any
program I know of today excepting Deep Blue, based on things like the
Win At Chess results I have posted here before.  It uses null-move, R=1,
non-recursive (only one null move in any single path).


If you'd have had the opportunity to sit around the table with Hsu and
company at any of the recent ACM events, you'd know his opinion of null-
move searching.  However, I should add, two years ago he thought my
pruning in the q-search in Cray Blitz and Crafty was wrong, but then in
DB-II he added such an idea.  So you never know, although I can't see him
doing null-move...  That is directly opposed to the extensions they try to
do.



>
>Jonathan Baxter



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.