Author: Uri Blass
Date: 09:12:48 06/06/03
Go up one level in this thread
On June 06, 2003 at 10:16:31, Thorsten Czub wrote: >On June 06, 2003 at 07:38:56, Uri Blass wrote: > >>I did not know that botvinik talked about it. > >Botwinnik's book PIONEER is very concrete in fact it has positions, >search trees and explanation. we should try to scan that book and make it public >on the net for the programmers. > >might be a good idea ! > >there is much said about botwinnik. most stuff comes from people >who never read the book. I did not know that he wrote a book about it. What I knew is only an interview with him when he described ideas in a way that cannot be translated to a chess program. > >>I read some things that botvinik said but it sounds to me too general to use >>them for a chess program. >>He never gave exact definition how to do things. > > >IMO the book is precise enough. > > >>I am surprised if botvinik's ideas helped you because the ideas that you talk >>about seem to me natural to think about them without botvinik. > >:-)) > > >>There are also other trivial ideas that are not implemented in most chess >>programs(for example the fact that programs has no knowledge about the second >>best move cause to absurd decisions by programs when they can use time for their >>move inspite of the fact that it is obvious that there is only one legal move to >>prevent mate). >> >>For me the main problem is to implement things without bugs and not to think >>about them and extending plans or calculating the exact value of the second best >>move at small depthes to decide about better time management are 2 ideas that >>were never implemented until today in movei because I decided that there are >>more important or more interesting things to work about them. >> >>Uri > >IMO you should open your horizont for ideas uri. > >since many chess programs have no idea about >PLANNING in chess, you can kill them if YOUR chess program can do so. I will not give details here in order not to help my opponents but one of the things that I hope to do is to make the illusion that movei plans and not only by searching lines that follow a plan deeper(something that today I do not do). I already do something small to encourage it and I believe that it can be improved significantly. It prevents me to use hash tables for pruning in the way that all programs do but I believe that the possible advantages are bigger. Uri
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.