Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Test position for handhelds (too easy for fast PCs)

Author: Uri Blass

Date: 02:09:32 04/01/02

Go up one level in this thread


On April 01, 2002 at 04:24:15, Uri Blass wrote:

>On March 31, 2002 at 16:32:22, Kurt Utzinger wrote:
>
>>On March 31, 2002 at 14:16:58, Roy Eassa wrote:
>>
>>>
>>>How long until it sees that White has a forced mate?:
>>>
>>>[d] 1k1K4/1p5P/1P6/8/8/8/p7/8 w - - 0 1
>>>
>>
>>Chess Genius 1.4 for Palm (16 MHz) does not find the solution with 7 min. And
>>even after 1.h8Q a1Q the program shows no mate within the next 9 min.
>>Kurt
>
>
>My program is even worse at that position
>
>It get stuck at depth 9 in the original position and after
>1.h8Q a1Q 2.Qg8 Qa2 it gets stuck after finishing depth 5
>and failing high at depth 6.
>
>The branching factor at depth 6 is clearly more than 1000.
>
>I suspect that the branching factor is an infinite number and
>I will have to investigate what is the problem.
>
>Uri

In order to investigate the problem
I asked my program to print it's main line every 2^20 nodes.

It seems that the problem is not infinite loop but
some extensions that get out of control in that position

I see that my program always print long lines of 21-22 plies
here thanks to some extensions.

I expect the next version not to have the same problems because
I am not going to let these extensions in this case
and similiar cases.

Uri



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.