Author: Steve Glanzfeld
Date: 16:53:13 07/25/04
Go up one level in this thread
On July 25, 2004 at 19:25:07, Jorge Pichard wrote: >On July 25, 2004 at 18:21:16, Steve Glanzfeld wrote: >>When you want to compare books, you must use THE SAME ENGINE for both, and play >>very many games. > >It doesn't make sense to play Shredder8 versus Shredder8 or List 5.12 vs List >5.12, also I don't need to play 1000 games if the difference is noticeable >within 20 games. It makes sense WHEN YOU WANT TO COMPARE BOOKS (see above?). When you play 20 games engine A + book X versus engine B + book Y, you can't tell anything about the difference of the books. I guess, that way not even 100 games would be sufficient. That would be useless. When testing component I., component II.,III. etc. should be equal or identical whenever possible. Isn't that pretty obvious? :) Steve >>Also, people should take a close look at the book options each, as they are >>saved with the Arena books (unlike Fritz). The options may not always be set as >>the user likes it, or thinks how they should be. That may have more influence >>than the book's quality itself has, on the opening performance. (And don't ignore this either :))
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.