Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: BWTC #5

Author: Bruce Moreland

Date: 12:53:43 08/22/00

Go up one level in this thread


On August 22, 2000 at 15:42:12, Brian McKinley wrote:

>I am testing my program against the BWTC suite of positions.  Because of its
>name I assume there is a mate for each position.  Although there are still over
>100 positions for which my program does not find a mate, I can usually see the
>mate myself.  I am having problems finding a mate for number 5.
>
>r2qk1r1/p4p2/bp2pQp1/1n1pP1Bp/7P/3P2N1/P1R2PP1/2R3K1 w q -
>
>Can anyone confirm this, and if so, do you know where there is a list of
>positions for this suite that don't have a mating solution?

The author of the study apparently assumed that after 1. Rc8 Rxc8 2. Rxc8, black
would either play 2. ... Bxc8 3. Qxd8#, or 2. ... Qxd8 3. Qe7#.

But the author didn't count on 2. ... Kd7.  That's a queen down but it's not an
immediate mate.

BWTC is so easy that it's not often used.  It's almost certainly full of little
mistakes and other problems.  If you have a program that's beyond the very early
stages, you should get all or almost all of the problem right in a fraction of a
second.

BWTC and WAC (Win at Chess) were both written by Fred Reinfeld.  The books
aren't very scientific, and clearly weren't meant to be used to test computers.
There's one position in one of those two books, where there are two ways to
mate, and Reinfeld specifically rules out one of them because it isn't as pretty
as the other one, even though they take the same number of moves.

bruce



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.