Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: shortest mate test position that programs cannot find

Author: Terry McCracken

Date: 11:53:34 07/20/02

Go up one level in this thread


On July 20, 2002 at 14:46:10, David Dahlem wrote:

>On July 20, 2002 at 14:11:30, Terry McCracken wrote:
>
>>On July 20, 2002 at 10:32:25, David Dahlem wrote:
>>
>>>On July 20, 2002 at 04:12:37, Uri Blass wrote:
>>>
>>>>On July 19, 2002 at 23:07:20, allan johnson wrote:
>>>>
>>>>>On July 19, 2002 at 23:05:58, David Dahlem wrote:
>>>>>
>>>>>>On July 19, 2002 at 21:23:19, K. Burcham wrote:
>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>what is the shortest "mate in x moves" position that you know of that programs
>>>>>>>have a diffcult time finding or cannot find at all? Please, not e4.
>>>>>>>
>>>>>>>will you please post the position?
>>>>>>>kburcham
>>>>>>
>>>>>>I tested some programs on this mate in 1 and some didn't see it. I don't
>>>>>>remember now which engines couldn't find it.
>>>>>>[D]5K2/8/2qk4/2nPp3/3r4/6B1/B7/3R4 w - e6 0 1
>>>>>>Regards
>>>>>>Dave
>>>>
>>>>I think that this is not the fair way to test
>>>>
>>>>You should give programs the following position
>>>>
>>>>[D]5K2/4p3/2qk4/2nP4/3r4/6B1/B7/3R4 b - - 0 1
>>>>
>>>>play e5 and see if they cannot see the mate in 1.
>>>>
>>>>In other case you use a bug that programs do not
>>>>support Fen correctly.
>>>>
>>>>My latest Movei still ignores the en passant in supporting Fen
>>>>so it is unfair to say that it cannot see the mate in 1.
>>>>
>>>>Uri
>>>
>>>The poster asked for a test position and that's what i posted. So whether it's
>>>fair or not is not revelant.
>>>Regards
>>>Dave
>>
>>Catch 22:o) If you use Uri's position, then 1...Kd7 avoiding mate.
>>If you give Dave's position to the computer, it won't understand there is an
>>en passant, as the move e7-e5 isn't in the computers history.
>>
>>So all you can do is give Dave's position, or show both, and mention that e7-e5
>>has been or will be played by Black.
>>
>>Unless the FEN can tell the program e7-e5 has just been played which allows for
>>en passant?
>>
>5K2/8/2qk4/2nPp3/3r4/6B1/B7/3R4 w - e6 0 1
>Doesn't the "e6" in the FEN tell the program that e7-e5 has just been played? I
>am no expert on FEN protocol, but some programs DO solve this.
>Regards
>Dave

I'm no expert of FEN protocol either, and if some programmes do solve this with
5K2/8/2qk4/2nPp3/3r4/6B1/B7/3R4 w - e6 0 1, then some _must_ understand _e6_
which would mean en passant.

Terry

>
>>If the computer can't know whether en passant is legal, it can't solve the
>>problem.
>>
>>Terry
>>
>>P.S. How many humans might forget the idea of en passant?;-)



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.