Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: not using nullmove?

Author: scott farrell

Date: 14:52:45 02/14/04

Go up one level in this thread


On February 14, 2004 at 05:28:04, Uri Blass wrote:

>On February 14, 2004 at 02:09:38, Dann Corbit wrote:
>
>>On February 13, 2004 at 12:51:08, Uri Blass wrote:
>>
>>>On February 13, 2004 at 12:41:41, Tord Romstad wrote:
>>>
>>>>On February 13, 2004 at 12:28:26, Mike S. wrote:
>>>>
>>>>>On February 13, 2004 at 11:07:02, Tord Romstad wrote:
>>>>>
>>>>>>(...) That several of the top commercial
>>>>>>programs do not use null move (at least not in the conventional way) is
>>>>>>further evidence that it is possible to come up with something better.
>>>>>
>>>>>Are you sure that there are top engines which don't use nullmove? Maybe they
>>>>>"just" combine it with some zugzwang detection, or switch it off depending on
>>>>>the amount of material earlier than other engines...
>>>>>
>>>>>(I think the last engine not using nullmove which was among the top was Chess
>>>>>Genius, but that was many years ago.)
>>>>>
>>>>>I'm aware that some engines are not affected by the zugzwang/nullmove problem as
>>>>>much as others are. - I'm only asking from the viewpoint of a user and fan, IOW.
>>>>>no programmer: Is it ok to say, nowadays all top-10 (if not more) engines have
>>>>>to have nullmove implemented in some way, because without they wouldn't be
>>>>>competitive in terms of search speed?
>>>>
>>>>I guess all strong programs use what Cristophe once called "the null move
>>>>observation", which consists of the simple fact that in almost all chess
>>>>positions, there is at least one move which is better than doing nothing.
>>>>Hence, in a wide sense of the term, all programs do indeed use nullmove.
>>>>
>>>>However, everybody does not use standard recursive null move pruning in
>>>>the form popluarized by Chrilly Donninger in the ICCA Journal a few years
>>>>ago.  Chess Genius and Junior have already been mentioned.  Two other
>>>>examples are Rebel and Chess Tiger.
>>>>
>>>>Tord
>>>
>>>Rebel and chess tiger use null move pruning.
>>>They use other reductions but the same also for you.
>>
>>At one time, Rebel did not use null move pruning (IIRC).
>
>At one time when I began with movei movei also did not use null move pruning.
>The point is that Ed found that adding null move pruning was productive.
>
>Uri
I read a post recently that Ed uses, but not like we do.

He uses it to confirm a reduction or pruning.

He uses it around alpha and not beta, and doesnt use it to cut off.

Scott



This page took 0.03 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.