Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: One mate to solve...(proof number search results)

Author: Angrim

Date: 22:21:19 05/16/01

Go up one level in this thread


On May 16, 2001 at 15:48:02, Dann Corbit wrote:

>On May 16, 2001 at 01:49:08, Angrim wrote:
>
>>On May 15, 2001 at 06:30:34, leonid wrote:
>>
>>>Hi!
>>>
>>>If you would like to solve one simple mate then you have one.
>>>
>>>[D]1qbBBbq1/1Q5r/R2Nq1P1/2Q1N1n1/KQ2q1rk/2Q1N1n1/R2Nq1P1/1Q5r w - -
>>>
>>>Please indicate your result.
>>>
>>>Thanks,
>>>Leonid
>>
>>proved that move b7xh7 wins, 9 turns
>>PN2:29753935 evals, 697015 expands, 306.22 seconds
>>
>>pv stub: b7xh7 g8xh7 d6f5 h4h5 f5xg3
>
>That's very impressive!
>Can your program find the other checkmate in the same distance as well?

no.  I don't have an option to delete the winning move from the list and
resume searching, which is what would be needed.
It actually does much better in suicide chess than in regular
chess, to the point that most of the stronger suicide chess engines
now use a short pn-search to check for wins before the main minmax
search starts.

Angrim



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.