Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Storing AB Bounds for mate scores... what is correct?

Author: John Boyd

Date: 22:51:53 05/13/03

Go up one level in this thread


On May 13, 2003 at 22:34:57, Robert Hyatt wrote:

>On May 13, 2003 at 21:30:01, John Boyd wrote:
>
>>Hi all,
>>
>>Usually, a mate score is greater than beta and thus is stored as a LOW BOUND in
>>the hash table.
>>
>>But why? Assuming its the true score, shouldn't it be stored as an EXACT score?
>>
>>I guess the problem is in assuming it is indeed the shortest mate. Is that why?
>>
>>Thanks,
>>
>>Ross
>
>you can have exact mate scores, _and_ mate bounds.  They both can be stored
>just fine if you don't have bugs.  The main point is to adjust a mate score/
>bound so that it is mate in N from the current ply, rather than mate in M
>from the root...

Thanks for the feedback. Last week I did the mate score fix you describe and now
its returning the correct scores on subsequent searches. That's 1 problem
solved... and 999,999 problems to go...





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.