Computer Chess Club Archives




Subject: Re: Mate that every program will solve...

Author: leonid

Date: 10:12:22 04/17/01

Go up one level in this thread

On April 17, 2001 at 11:13:02, Heiner Marxen wrote:

>On April 17, 2001 at 07:43:30, leonid wrote:
>>If you like to solve very "regular" mate position, then you have one.
>>[D]R7/4P1Pq/1rR5/2pb1rnk/1Q3B2/4Nn2/1PB2Np1/2K1b1Q1 w - -
>>Please, indicate your result.
>According to Chest there are two key moves for mate in 7 moves:
>Qxg2 Bd2+ Qxd2   Bxc6 g8=Q   Qxg8  Rxg8   Bd5  e8=Q+  Bf7  Nxf5   Nd4  Rxg5#
>Rh8  Bd2+ Qxd2   Be4  e8=Q+  Rf7   Qxf7+  Nxf7 Qd5+   Bf5  Rxh7+  Nh6  Rhxh6#
>(K7/600, 350MB hash, 9.3 minutes)

Thanks, Heiner! My program demonstrate the worst possible branching factor for
this position. Even for seeing the 6 moves by brute force it took already 38 min
(Celeron 600. No hash). Only after your indication I went to see  7 moves by
brute force. It took for mine hefly 1 hours 5 min. And this is after my last
improvement that cut 6 moves brute force search already by 48%.

Previously this position my solved by selective search in 10 moves.


This page took 0.21 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.