Author: leonid
Date: 10:05:26 05/08/01
Go up one level in this thread
On May 08, 2001 at 10:56:31, Heiner Marxen wrote: >On May 07, 2001 at 17:48:03, Paul wrote: > >>On May 07, 2001 at 16:23:20, leonid wrote: >> >>>Hello! >>> >>>And how about solving mate in two variations? We will start with female side. >>>Contain more esthetic and demand very gentle approach. >>> >>>[D]Qn1q1nQ1/1QNbNQ2/1qqkqq2/1QrqrQ2/1q1N1q2/3q4/B1RbR1B1/B2K2B1 w - - >> >>Hullo! >> >>Mate in 8 in half a minute ... >> >>00:27 WM8 06 Ne8+ Qxe8 Nc8+ Qexc8 Qgxf8+ Qfe7 Q7xe6+ Rxe6 Rxe6+ Qxe6 Qxe6+ Bxe6 >>Qbxe7+ Ke5 Nxc6# > >Chest find two solutions (in 8 moves (!), i.e. Pretz is shortest, again :-) > >Nc8+ Qxc8 Ne8+ Qcxe8 Qaxb8+ Qbc7 Q7xc6+ Qxc6 Qxc6+ Rxc6 Rxc6+ Bxc6 Nb5+ >=*= Qfxc7# > >Ne8+ Qxe8 Nc8+ Qcxc8 Qgxf8+ Qfe7 Q7xe7+ Q6xe7 Qaxb8+ Qbc7 Qxd5+ =*= Qxd5+ >Rxd5 Qxd5# > >(K7/600, 350 MB hash, 517 seconds = 8.6 minutes) > >The second one is not yet ready. > >Cheers, >Heiner Hi, Heiner! Today had the chance to think about what you explained for mate solver logic. Especially where you indicate that you take in consideration captures value in order of aligning all moves in the chain. I have something like this in second part of my program. There all moves, after move generator, are aligned after "best previous move" and value of each possible capture. Simply move with capture of queen qill go ahead of move with capture of rook and so like. Put this procedure instead of one specially written for mate solver. Result is still not clear but it look like promissing to my surprise. Only big number of positions will say if this will work. Even if both procedures (new for mate solver and old for the "other part of chess program") are identical in speed, I will be happy. It could signify that I could cut from my code new one and use only old. Leonid.
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.