Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: One mate to solve.

Author: leonid

Date: 05:46:06 09/01/01

Go up one level in this thread


On August 31, 2001 at 23:50:50, Angrim wrote:

>On August 31, 2001 at 06:16:57, leonid wrote:
>
>>[D]2r2k2/2pqqr2/1nq2bb1/qnQQQQqq/q1Q2Q1q/2Q2Q2/2R2R2/2B2K2 w - -
>>
>>Please indicate your result.
>>
>>Thanks,
>>Leonid.
>
>older pn2 code:
>proved that move e5xe7 wins, 9 turns
>PN2:3496323 evals, 74810 expands, 17.28 seconds
>
>newest pn2 code:
>proved that move c5xe7 wins, 9 turns
>PN2:14645477 evals, 361828 expands, 84.17 seconds
>
>newest pn-search:
>proved that move e5xe7 wins, 9 turns
>PN:1810301 evals, 41852 expands, 10.39 seconds
>
>one downside to selective searchers: very unstable behavior.  In
>all cases the search settled on a favorite move in the first second
>or so, then spent most of the remaining time proving that it won.
>It seems that e5xe7 is simpler to prove than c5xe7, so the 5:1 speed
>difference may well have resulted from a change in move ordering.

Hi, Angrim!

As I see you found three solutions and all (if I read correctly) 10 moves deep.
It is actually mate in 10 and you  found mate in shortest way. I proved that
position is mate in 10 by searching the first 9 moves by brute force and mate
move by 10 moves deep.

Your time is excellent, for sure. Mine was 15 seconds for selective, for first
mate. Celeron 600. No hash.

Cheers,
Leonid.

>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.