Author: Robert Hyatt
Date: 17:49:21 05/08/04
Go up one level in this thread
On May 08, 2004 at 17:54:50, Uri Blass wrote: >I would like to read some descreption of the book algorithm of different free >programs. > >I am still thinking about the book structure of movei >I do not like the solution of Crafty of making all the possible moves and trying >to find if one of them lead to a book position. > >I think that a better solution may be to have some books: > >book A will include mainly position move information(I think that the decision >should be also based on the time control) > >book B will include mainly position that the program likes and scores for them >and may be updated after every game. > >book C will include positions and their parents and after updating book B may be >used to check which positions should be updated in book A. > >I also think about the learning problem and a possible problem is that a move >may be bad for a program at blitz but good for it at longer time control so I am >interested to know what programs with learning function do in order not to learn >misleading information based on blitz games about longer time control. Go to my web site and look at the book learning paper. It gives lots of ideas about various problems and solutions... > > >I will probably add no additional learning in the first version that includes >book that know about tranpositions and I will only add book in position move >format but I would like to know what is the decision that other amateurs do >about book and learning. > >Uri
This page took 0.01 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.