Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Verified Null-Move Pruning, ICGA 25(3)

Author: Uri Blass

Date: 19:15:39 11/21/02

Go up one level in this thread


On November 21, 2002 at 22:05:15, Vincent Diepeveen wrote:

>On November 21, 2002 at 11:14:24, Robert Hyatt wrote:
>
>Just go for a 2 minute single cpu implementation.
>
>But of course you will find that R=2 always will outperform
>already his verification search with 1 ply reduction.
>
>You can implement it quicker his algorithm by doing next
>after nullmove:
>
>int Search(alfa,beta,depthleft,allowingnull) {
>  ..
>
>  if( !allowingnull && nullmoveok ) {
>    .. try nullmove
>    if( score >= beta ) { // score returned by nullmove
>      score = Search(alfa,beta,depthleft-1,true);
>       // last parameter to prevent another reduction next position
>      if( score >= beta ) {
>        StoreInHashtableAndKillerTable(..);
>        return score;
>      }
>    }
>  }
>  ..
>
>This is a more elegant way to implement it, and it was already done
>years ago and published in ICCA journal.


No

It is not his algorithm.
The verify parameter in his article is not allowingnull.

It is a parameter that tells the computer if to allow searching to depth-1 in
the subtree.

Uri



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.