Author: Robert Hyatt
Date: 21:21:16 01/07/00
Go up one level in this thread
On January 06, 2000 at 21:29:57, Mike S. wrote: >On January 05, 2000 at 22:12:45, vincent dichiacchio wrote: > >>One of my chess programs always transposes to books when it can, even when it >>shouldnt. > >What do you mean, are the book move worse than what the program would calculate >itself? In this case the book must be bad, or not in tournament setting. >In the Fritz GUI's, it is a user setting if the book should be checked always or >not. But normally, transposing back as you say, should be an *advantage* when >the book is well done, consisting of reliable theory moves. >Some opening book, especially large one's generated from game databases, have >been critizised because of blunders, but also because they may be *too large*, >delaying the engine start to a point where the middlegame may be already shaped, >leaving not enough "game" for the program. > >Regards, >M.Scheidl There are some nonsensical transpositions: 1. e4 e5 2. Bb5 a6 3. Nf3 and now it is black's move. If he plays Nc6 he transposes back into a well-known Ruy Lopez line. If he plays axb5 he wins the game instantly. :) 2. Another bad thing. You have a bad line that your program won't play, because a particular move is marked as ? to prevent this. So you search and try to find something else, and do. Your opponent makes the next book line move anyway and _now_ your program transposes right back into book by playing the move that was marked with a ?, since it is now one move further into the line and this move is not really a known book line except for the transposition. So you transpose back into a line you know is bad. And you lose again. This has happened to me a few times... 1. I have pretty well solved. 2 is much harder.
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.