Author: Heiner Marxen
Date: 13:48:02 09/14/99
Go up one level in this thread
On September 13, 1999 at 18:17:35, Dann Corbit wrote: >That is a fascinating problem. I intend to pour over the output until I >completely understand it. > >Crafty never did pick the key move, but did stop assuming that the offered rook >would be taken: >8/8/8/1k3p2/p1p1pPp1/PpPpP1Pp/1P1P3P/QNK2NRR w - - acd 27; ce 2283; pv Rg2 Kc6 >Kd1 Kd6 Ke1 Kc6 Re2 Kd6 Kf2 Kc6 Rg1 Kd5 Ke1 Kd6; > >CHEST did not find any mate through 16 fullmoves, and hit the wall trying to >calculate 17 fullmoves. [Something powerfully exponential must happen at that >point, since it pounded through 16 moves in a short time, and failed to >complete 17 moves over an entire weekend.] Yes, CHEST starts to detect that capturing the white pieces leads to mates. CHEST very much likes to capture the attackers pieces, hoping to cut down the size of the tree. With depth 15 CHEST does already find some partial mates in 9, and needs up to 3 (or more) answers (black moves) to find one, which avoids a mate. That is exceptionally bad. CHEST is fooled, here. :-( >The assessment that computers will never solve this sort of thing is probably >an exaggeration, but *currently* it appears to be beyond the limits of what >a PC can do (at least). Yes. A program on a PC solving this problem would be a major breakthrough, IMNSHO. CHEST simply can't do it (currently). Don't hold your breath. Anyhow, this is a very fascinating problem, and I will study it a bit more. Heiner
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.