Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Question on Probing Hash Table for Best Move in QSearch

Author: Robert Hyatt

Date: 07:50:23 12/07/04

Go up one level in this thread


On December 06, 2004 at 21:52:40, William Bryant wrote:

>On December 06, 2004 at 21:35:54, Robert Hyatt wrote:
>
>>On December 06, 2004 at 20:48:57, William Bryant wrote:
>>
>>>Have other people tried to probed the hash table in the QSearch for a best move
>>>to improve
>>>move ordering.
>>>
>>>I realize, that if it is not a capture or promotion, it will be dropped.
>>>
>>>But if it is a capture or promotion, search it first.
>>>
>>>Note: I am not talking about storing best moves, just checking for them.
>>>
>>>
>>>Any thought or discussion would be appreciated.
>>>
>>>
>>>William
>>
>>
>>It will work just fine, but whether it is worth the trouble or not needs
>>testing.  IE a capture might look good if you can follow it up with only a
>>capture, it might look worse if you can play a non-capture.  This means that the
>>move might look good near the root, but bad in the q-search where it fails to
>>produce a cutoff as expected...
>>
>>Needs lots of testing.  And don't forget your obvious mention that you have to
>>be sure the best move really is a capture or you will blow your tree up...
>
>Thank you,
>
>On my first attempt at an engine,
>I ultimately took the inCheck test out of the QSearch with an overall speed up.
>
>What is the current thinking on doing a full width search in QSearch if you are
>in check?
>
>William

About the same as any other "mostly-opinion-based" questions in CC.  Some do,
some don't, there are good and bad reasons for doing it either way. :)

I have a very simple q-search, for example..




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.