Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: null-move Q-search

Author: Robert Hyatt

Date: 18:47:54 10/30/99

Go up one level in this thread


On October 30, 1999 at 20:48:19, Dave Gomboc wrote:

>On October 30, 1999 at 20:30:24, Robert Hyatt wrote:
>
>>On October 30, 1999 at 19:43:49, Bruce Moreland wrote:
>>
>>>On October 30, 1999 at 17:11:49, Christophe Theron wrote:
>>>
>>>>On October 30, 1999 at 16:15:05, Bert van den Akker wrote:
>>>>
>>>>>Can sombody tell me how the null-move concept can be used in
>>>>>the Q-search.
>>>>>
>>>>>Thanks in advance
>>>>>
>>>>>Bert van den Akker
>>>>
>>>>I was wondering too, but I didn't dare to ask! :)
>>>
>>>It is the normal standard quiescent search.
>>>
>>>In your quiescent search, you compare the score of the position if you don't
>>>move, with the score of various capturing moves, and return whatever is higher.
>>>
>>>bruce
>>
>>
>>I don't think that is what Beal did.  He would try a real null move and let the
>>other side make another capture, then use that value to establish an alpha,
>>rather than a normal call to evaluate().  When alpha met beta, the search would
>>not progress any further.
>>
>>At least that is my memory of the thing from a _long_ time ago.  I have the
>>ACC book at the office and will peek again Monday.
>
>Do you remember the ACC volume?  I have a stack of them in my office at school,
>and I will be there on Sunday.  Maybe I can just flip through them...
>
>Dave


I am not sure.  I have a 'review' copy that I was asked to proofread.  And I
think the one I reviewed went into one of the Advances in Computer Chess
books that Beal used to edit after the Advances in Computer Chess conference
he would orgranize on a regular basis.

I'll try to sort thru everything monday and find a real citation for it that
you can find...



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.