Author: Bruce Moreland
Date: 11:52:26 04/22/00
Go up one level in this thread
On April 22, 2000 at 10:51:27, Pierre Bourget wrote: >Is it true in this position that Hiarcs play Qf7+ ?? because of the tablebase ? >Isn't it a little ridiculous.Look here for Tim Krabbe opinion: > >http://www.xs4all.nl/~timkr/chess2/diary.htm > >[D]5k2/2Q5/8/8/5p2/5P2/5P2/K7 w - - 0 1 I like Tim Krabbe's stuff a lot. What he's complaining about here is a transition case. Some programs probe endgame databases at the root, and if there is a mate available, they select it. Other programs probe endgame databases at the tips, but once they determine that there is a mate, they stop searching and move immediately. Both of these techniques would result in rapid play by the computer, but you would see the sub-optimal kind of play that Mr. Krabbe is complaining about. If you probe at the tips you can also do a normal search, which would avoid the problem, but would lead Mr. Krabbe to complain that a program searches three minutes per move in a KQ vs K ending. There are ways to smooth over these problems, but it is difficult to solve the problem perfectly (meaning, it is difficult to simulate human play, for good or bad), and every change introduces the possibility of adding a bug, which could turn an absolutely won game into a draw or even a loss. If the goal of the game was to play this part of the game in a certain style, it would be easier to fix this problem, since you'd have something specific to shoot for. But these programs are generalists, and they have been programmed to win in whatever manner they can. That is ultimately why these problems are not fixed. Once the fish has taken the bait, a human fisherman has to contend with the strength of his fishing line, and with the possibilty of getting water down his boots, so he engages in technique that is as artistic as what he was doing before the fish became hooked. A computer program can simply use a 1/4" steel cable as fishing line, and a power winch. The only problem is remembering not to pull the fish in so fast that you rip its head off. Not artistic at all, but extremely effective. These programs have been told to be effective before everything else. 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.