Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Null Move Heuristic - Comments Please (somewhat long)

Author: James Swafford

Date: 21:47:12 05/14/99

Go up one level in this thread


On May 15, 1999 at 00:34:35, Dave Gomboc wrote:

>On May 14, 1999 at 20:10:09, William Bryant wrote:
>
>>I'm ready to ruin a perfectly good search by adding more code. If done right,
>>the Null move should make it much stronger.  My copy of C. Donningers paper has
>>been ordered by inter library loan so I have not yet read the paper.  Below is a
>>summary or digest of information available from this forum over the last year
>>concerning null moves.
>>
>>I think this is a pretty up to date summary of the heuristic, but would
>>appreaciate any comments.
>>
>>Note: I have assembled this from many posts, maybe even yours.  I claim no
>>copyright on the material.  For those, like me, ready to expand their programs
>>with Null move search, I hope this summary helps.  Please feel free to correct
>>any errors I might have made.
>>
>>William
>>wbryant@ix.netcom.com
>>
>>Null Move Summary
>>
>>Description
>>	Null moves are a forward pruning mechanism to generate a beta cutoff without
>>doing a full search
>>
>>Situations to Avoid Null Moves
>>1. When the side on move is in check. Then a null move simply allows the other
>>side to capture the King.
>>2. Having just done a null move.
>
>I have heard this restriction before, but I have also heard that it is fine to
>do this, because one or the other null-move will fail.  Does this make any
>sense?  I have not spent much time thinking about it.
>
>Dave


Suppose you're in Search( ) with depth remaining=7.
Now you start a null search w/ a reduction of 2, so
you're in search again with depth=5.

Now repeat.  Repeat again.
Soon enough you're in quiescence, and you might as well
have set r to 7.

--
James




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.