Author: Ed Schröder
Date: 10:59:49 05/14/98
Go up one level in this thread
>Posted by Amir Ban on May 14, 1998 at 09:54:26: >In Reply to: why do computers waste time about forced moves? posted by blass >uri on May 14, 1998 at 09:02:54: >On May 14, 1998 at 09:02:54, blass uri wrote: >>when I give junior to play with fritz I see both programs >>waste time about forced moves(there is more than one legal move >>but only one logical move to prevent checkmate. >>I suggest that the program will use in calculating one second to >>compute the difference between the best move and the second >>best move. >>if the second best move and decide about maximal time >>that it should consider the move. >>It is only a small improvement >>because most of the moves are not forced >>but this improvement seem to me simple to program. >>if the next best move give the opponent >>the possibility to do a checkmate the program should use only 1 >>second calculating the move. >>if the second best move lose a knight it should use more than >>1 second(because maybe it is a mistake) but not much time. >>I suggest the maximal time can be by the formula 1+(c/x) >>when c is a constant and x is the difference between the best move >>and the second best move (if the difference is more than a pawn) >It's a mistake to play "obvious" moves too fast. Rebel does that, and I >believe it lost quite a few games on this account. It was a customers wish from the Mephisto time. Bottom line, better make a bad move so now and then (< 1%) then waiting 10 minutes for a simple recapture. Like "selective search" it will never be equal as "brute force", still the final result is better. This time the result is not better for the playing strength of Rebel but it's there for the satisfaction of the customer. It's easy to remove the "quick capture" algorithm and gain some (5?) elo points but I think it's not worthy the complaints afterwards. - Ed - >I already do something along the lines you describe. It can be improved, >that's for sure. I think that to find out that a move is "obvious" needs >some non-trivial investigation that most programmers, realizing it's not >too common, do not care to invest. >Amir
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.