Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Null move generalization

Author: Dann Corbit

Date: 13:40:01 04/17/02

Go up one level in this thread


On April 17, 2002 at 16:32:47, Uri Blass wrote:

>On April 17, 2002 at 16:09:27, Dann Corbit wrote:
>
>>On April 17, 2002 at 15:43:23, Jesus de la Villa wrote:
>>
>>>
>>>Have someone defined the general rule(s) where null move
>>>is unable to find simple combinations?, and if so, which
>>>are those rules ?
>>>
>>>"Obviusly" is more expensive to check it than to not
>>>use Null Move.
>>
>>There are times you simply must turn null move off or disasters will happen.
>>
>>When the board gets sparse, turn it off.
>>
>>If you are in check, turn it off.
>>
>>If you are already searching with null move, additional null move pruning [for
>>subsequent plies] is questionable.
>>
>>If you see a checkmate threat during the null move search, turn it off or
>>extend.
>>
>>Null move will not remove the ability to find simple combinations.  It only
>>delays it.  But it might delay it enough that you will not find it in a
>>reasonable time.
>
>Null move may removes the ability to find simple combinations when zunzwang is
>involved.
>
>It happens to a lot of programs that do not use zunzwang detection(I think that
>crafty is one of them and at least it was the case in the latest version that I
>checked).

It only reduces the depth of the search.  So (for instance) a tactic that might
take 7 ply to find with a full-width search may take a 9 or 10 ply search to
find with null move turned on (depending on the value for R).

I don't see how it can completely remove the tactic from ever being seen unless
the implementation of null move is broken.



This page took 0.01 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.