Author: Uri Blass
Date: 10:03:51 08/25/02
Go up one level in this thread
On August 25, 2002 at 11:16:56, Mike S. wrote: >On August 25, 2002 at 10:12:53, Uri Blass wrote: > >>(...) >>It is important to get positions that you understand from the opening. >>If the fritz book is leading your engine to positions that your engine does not >>know to play then it is a disadvantage. > >So you probably think movei would score *less* with the Fritz 7 book? :o)) No I did not say it. I am also more interested in teaching movei to find good opening moves by itself and not in learning book moves from theory so I do not care to waste time about writing book lines for movei(at least not in the near future). > >Book depth and variety (knowledge of seldom openings) matter *much* more. The >Fritz7.ctg is prepared for virtually *everything* IMO. I've seen very seldom >lines, gambit variations which are not played in Grandmaster chess, etc. > >A good book has approx. 7969878 phantastillions of positions where the >variations end (and still 2495898 phantastillions if you limit the book variety >in a reasonable way), so what's that talk about "understanding"? You can try to >tune some *main lines*... but then the opponent plays a seldom side-variation, >and you can forget the book tuning. You can't avoid that, and you just can't >fine-tune a large book for every possible opening line of 907396720374034 >possibilities. Impossible. If I know that the engine scores better with c4 than with e4 based on hundreds of games I can put c4 and not e4 in the book. e4 may be also in some automutic generated book but the program is going to consider it only after losing a game with c4. The big automatic book can be used only to avoid losing the same opening twice and when there is no danger of losing the same opening twice the program can continue on it's own. 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.