Author: Jeroen Noomen
Date: 07:08:40 01/04/01
Go up one level in this thread
On January 04, 2001 at 09:28:06, Uri Blass wrote: Hi Uri, I am not going to put all moves that take Rebel out of book quickly, in my own book. There are several reasons to do so: 1. You can't put EVERYTHING in the book. If I put the move 3 ... a6 in the Caro-Kann in the book, John will play 3 ... h6. Or 3 ... e6. Or something else. 2. There is ALWAYS a way to take a program out of book quickly. As White I would suggest 1. e3 2. a3 and 3. d4 (already played in AEGON) as a reversed French. Or 1. d4 and 2. a3. Or 1. b3 and 2. a3. There are simply too many ways to get you out of book quickly. 3. Why should I spent months on such weird variations? If it MIGHT be played in 1 or 2 games versus humans? What is the purpose? 4. I concentrate on REAL openingbooks, novelties, popular lines and less popular lines. 5. If a computer program fails to make an advantage after such a move as 3 ... a6?! it proves that is not a good player. As long as computer programs do not play well against blocking positions, they are not of IGM strength. 6. So why should I mask the weakness of the computer program by putting EVERYTHING in the book? The problem is, that computer programs DO NOT KNOW HOW TO HANDLE THESE POSITIONS. That is the real problem! That is what programmers should concentrate on. Not an attitude like "I need not do something about anti-computer chess, since my opening book will prevent it. And if not, I will blame the opening book maker". So I will not put these silly lines in anyone's book. Because then I would prevent the REAL solving of the problem: Namely how to play against an anti computer strategy, a blocked position, a move taking it quickly out of book, a king's attack or whatever. It is in the interest of computer chess that we witness these kind of games, because then people start do something against it! I am sorry Uri, but I am NOT going to mask the problems of a computer program by filling all the 'holes' in a book. You should accept what the problem really is: That computers are unable to tackle anti-computer styles yet. THAT SHOULD BE CURED. Jeroen >I looked at the chessbase online database and found 13 games with the position >after 1.e4 c6 2.d4 d5 3.Nc3 a6 so Van der Wiel's move is not a novelty. > >How many positions do you need to include in Rebel's book in order to prevent >the opponent to get Rebel out of book without making a novelty. > >Rebel does not have to know all the positions in the games becausethe user book >prevent most of the position(if the user book has only 1.e4 or 1.d4 then games >with 1.c4 for the opponent are not relevant. > >I think that it may be productive if Ed writes a program to find all the >possible positions that can happen when Rebel is out of book not because of a >novelty. > >He may give these positions to you in order to help you to get a better opening >book. > >I guess that the number of the relevant positions is between 10,000 and 100,000 >and the number of relevant position when the position was played at least 13 >times is clearly smaller. > >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.