Author: Stefan Meyer-Kahlen
Date: 08:34:09 04/21/04
Go up one level in this thread
On April 21, 2004 at 09:49:31, William Penn wrote: > >I'm not sure if I made myself clear in my original post. I'm talking about... > >(1) For the sake of an example, let's say it is white's turn to move initially >in a game in progress which has been saved to disk. >(2) So the game is loaded from disk, and Shredder is set to calculate for a long >time (overnight) to try to find white's best move. I have always used Infinite >Analysis mode for this process. >(3) Then I press the spacebar, and Shredder makes the best move it found on the >chessboard. >(4) Now it is black's turn to move, and ideally Shredder should retain the >analysis obtained when white's move was calculated. It should not have to start >a calculation from "ground zero" again (at the beginning, with no knowledge). If >it must start at ground zero again, this is highly (absurdly) inefficient. Yes, doing this would be stupid. This is why Shredder is not doing this. Stefan >So that's what I'm talking about. > >I'm not talking about "positional learning" such as is stored in the >shredder.plr file for openings, etc. I'm talking about retention of prior >analysis to be used in making the next successive move(s) in the game. Perhaps >it all can't be retained, but most of it should be retained. > >Partly the reason for this is that we know Shredder's displayed analysis cannot >be trusted beyond the first move. So it must be iterated move by move in manual >fashion, if we want to find the true best line!? >WP
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.