Author: James T. Walker
Date: 18:59:23 05/27/99
Go up one level in this thread
On May 27, 1999 at 19:43:43, James Robertson wrote:
>One of my friends tried my program on this 'Mate in 2' from John Thursby's 75
>mate in 2 problems.
>
>FEN 8/8/3PQ3/2p2r2/2N2R2/2k2N2/3RK3/8/ w - -
>
>My program didn't see it, supplying a mate in 3 line (my email program crashed
>and I lost the message.) Crafty also gave a mate in 3 line, 1. Qe3+ Kb4 2. Rb2+
>Ka4 3. Qa8#.
>
>I responded with this, and he said no! It is still mate in 2, and supplied this
>line: 1.Na5! {threatens Rc4# or Qe3# or Qc4# or Qb3#} 1 ... Rxf4 2. Qb3#
>
>I have not had time to check on it... is it mate in 3 or 2?
>
>James
>
>P.S. He also supplied this funny mate in 2 links:
>http://misc.traveller.com/chess/graphics/cake.htm
>http://www.hsv.tis.net/chess/graphics/boi.htm
Hello James,
You will find this is very common and not a "Problem" as I see it. Many
programs will find a "longer" mate due to check extensions. Because the shorter
mates are not always all checks you may not find it until later if your program
does not stop on finding the "Long" mate. This particular mate though, should
only take a one ply search with check extensions. Probably your program
searches the queen move before the Knight move and therefore finds the mate by
moving the queen + check extensions before finding the shorter mate with the
Knight move + check extensions. I think you could avoid this by forcing your
program to search the full ply before accepting the first mate found. Then you
could use the "shortest" mate found.
Jim Walker
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.