Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: not using nullmove?

Author: Tord Romstad

Date: 01:33:15 02/14/04

Go up one level in this thread


On February 13, 2004 at 14:56:44, Anthony Cozzie wrote:

>If I understand Ed's homepage, he uses the standard recursive null move pruning
>everyone else does, except that he turns it off a low depths and switches to his
>static forward pruning.

That's what I used to think, too, but it turned out that I was wrong.  Take a
look at
the recent thread "I hate null moves" for the discussion between Ed and me about
Rebel's use of null moves.

There are some sound ideas behind the null move heuristic.  It is too difficult
and
risky to detect all deep threats by static methods, therefore it is useful to do
tactical verification searches with reduced depth and/or evaluation.  Allowing
the
same side to move several times in a row is also interesting, and can help us
find
plans as well as tactical threats.

However, I am convinced that there are better ways to combine the ideas of
reduced-depth searches and making several moves for the same side than
recursive null move pruning.  Recursive null move pruning is very simple and
primitive, and uses hardly any domain-specific knowledge.  It should be
possible to come up with techniques which do not suffer from the same
defects as null move pruning, and at the same time are even more effective
at reducing the tree size.

Tord




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.