Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Composed Problems Useful as Test Positions?

Author: Robert Henry Durrett

Date: 11:29:40 06/21/02

Go up one level in this thread


On June 21, 2002 at 14:19:21, Dann Corbit wrote:

>On June 21, 2002 at 09:56:27, Gian-Carlo Pascutto wrote:
>
>>I think the problem of most composed positions is that they
>>usually present very unusual positions which are exceptions to
>>a general rule and such.
>>
>>In practise, those positions won't happen, and the general
>>rule (which is the one that is in the chessprogram) will work.
>
>However, the crazy composed positions have two big advantages:
>1.  They are great fun (I wonder where leo has gone)?
>2.  They stress possibilities (e.g. your move generator cannot handle more than
>150 generated moves and a position has 218 so BOOM.  Better to find out now than
>in a WMCCC[1]).
>
>[1] Murphy's law *always* surfaces at these contests.  If you have a bad book
>line, it will surface.  If you have a serious bug, it will present itself.


Perhaps you could even carry this idea one step further:

Those engine designers who are lucky enough to also be strong chessplayers might
be sufficiently creative and innovative, chesswise, to come up with test
positions which are "customized" to test new subroutines, algorithms, coding, or
whatever.

The nice thing about difficult composed problems published in the open
literature is that they, perhaps, likely will present problems for the engine
which the programmer did not forsee.  They may be "rude surprises."

Are there any chess engine designers/developers who are also "chess problem
enthusiasts," proficient at composing and solving chess problems?

Bob D.



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.