Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Null move and move ordering stats

Author: Peter Kappler

Date: 15:45:38 12/20/01

Go up one level in this thread


On December 20, 2001 at 15:13:14, Ren Wu wrote:

>On December 20, 2001 at 01:45:37, Peter Kappler wrote:
>
>>
>>Here are Grok's numbers (full-width search only, not quiescence):
>>
>>null move, futility pruning:       77%
>>null move, NO futility pruning:    82%
>>NO null move, NO futility pruning: 85%
>>
>>I guess I need to do some work on my move ordering.
>>
>>-Peter
>
>Say if you change your move ordering code, and manager to increse this by a
>delta. Then question become
>
>Will this increase the program's strength?
>


Hi Ren,

Better move ordering means a smaller search tree, and thus a deeper search,
which always helps.  (Assuming that your improved move ordering code isn't too
slow!)

As was pointed out in other posts, this particular statistic favors engines that
use a SEE and/or internal iterative deepening to choose a good first move when
no move is available from the hashtable.  My program uses neither, and I suspect
that's why my numbers are so low.

-Peter






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.