Author: Robert Hyatt
Date: 06:10:20 01/26/99
Go up one level in this thread
On January 25, 1999 at 23:13:10, William Bryant wrote:
>On January 25, 1999 at 22:30:03, William Bryant wrote:
>
>>Shouldn't a mate-in-n score generate a cutoff and shorten the search
>>significantly.
>>
>>My new hash table seems to work well, however, in the end game (no table bases
>>yet), it finds mate-in-n but continues to search thousands and thousands of
>>nodes.
>>
>>Does a mate in n score need to be handled differently or is my hash code buggy.
>>
>
>
>Can this be handled inside the iterative deepening loop
>
>ie
>
>#define MateScore (9999 - MaxIterations)
>
>for (int depth = 0; depth < MaxIterations;depth++) {
>
> x=search_Root(initialAlpha, initialBeta, depth);
>
> if (x > MateScore) || (depth < -MateScore)
> break;
> //this is checkmate in the score backed up in the pv
>}
>
>or will this screw up bigtime somewhere.?
>
>Is there a better way to handle this?
>
>Thanks
>
>William
>wbryant@ix.netcom.com
I would ignore it... but you can handle it where you choose the 'next move'
to search. ie at the top, if (value >= MATE-300) return(NO_MORE);
so you don't search more moves at any ply where you have found a mate. You
have to do a little more than that to be sure you always 'make progress' but
that can work. But really, who cares?
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.