Author: William H Rogers
Date: 15:20:29 08/26/02
Go up one level in this thread
On August 26, 2002 at 17:34:17, José Carlos wrote: > I've explained how my program works regarding book here in the recent past >(search the archives for a more detailed information). In summary, my program: >checks what moves it can play out of book, determine which one is more >interesting according to some parameters (including some configurable randomnes, >of course) then decides if the move is clearly good or dubious or clearly bad. >Then makes it or starts searching. If it searches, it will try to avoid that bad >move if possible. In any case, it will remember how that search went, how the >following positions looked for the program and what what the final result of the >game. That data will allow the program a better choice next time. If it finds a >move itself that it considers better than book moves, that will be included in >the book as learned. > So my program is able to create its out book from nothing at all. Of course, >my program is not strong enough to make this feature interesting for the public, >and I've probably spent too much time in the book, but I'm happy with it. > > José C. Now Jose, I am begining to understand why you have been saying the things that you do. There are lots of other programs that add to and improve their books and some even start with nothing, but the majority of the top programs out there use proven books that were created from master games. Those that are used in this way just select the best move from their libraries as they already know that they are the best. Now if you assume that your program can select a better move than some of the old grand masters, that is if you are using any of the lines that were created by them, then you must have a pretty good program. If you are just creating your own books from whatever source then you are right in stating what you say, it is just that most programs do not operate that way. When they are using lines of openings created by old masters, they just make the predicted move and be done with it, and when they do that their heart of their engines never get accessed. They just play the 'quoted book move' and then wait for their opponent to make its reply. Once an opponent makes a move that is not in the programs book library, then it thinks or creates its own reply and once it decides which one to move, then it starts ponder just what its opponent might repy to the move it just made. Bill
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.