Author: Robert Hyatt
Date: 17:05:52 10/15/03
Go up one level in this thread
On October 14, 2003 at 18:02:26, Igor V. Korshunov wrote: >On October 14, 2003 at 10:08:47, Robert Hyatt wrote: > >>On October 14, 2003 at 08:00:48, emerson tan wrote: >> >>> >>>I played a 24 game match between chessmaster9000 default and chessmater9000 >>>without book with time control of 40 moves in 2 hours and the result was 13.5 to >>>10.5 in favor of cm without book. 8 wins for without book and 5 wins for the >>>default and 11 draws between them. >>> >>>Its seems that the one with no book "understands" the position better than the >>>one without the book because it calculates from move one therby can follow the >>>progress of the position from move one, the computer decides how the position >>>progress and it arives in a position that the computer likes and knows how to >>>play it. unlike the one with book,it may not know well how to play the position >>>after it is out of the opening. that's just my wild way of explaining it, but >>>seriously, why did the one without book beat the default one? I was expecting >>>the default one to beat the one without books because you can see that from the >>>earlt part of the game, the one with book was usually way ahead of of time, >>>leaving it much more time to calculate complications in later part of the game. >> >> >>Something has to be wrong. In a 40/2 game, most openings should go for 10 >>moves. That gives one program an extra 30 minutes on the clock since book >>moves are played instantly. That extra 30 minutes is significant. Add to >>that the fact that several openings have deep traps that can't be seen by a >>normal search, and the gap should widen farther. > >30 minutes is not significant. It is can give in average 10 Elo points. Such >difference can't be noticed in 24 games match. > >Finally, I can say that 24 games is not enough to decide that this result >strange. It might be more than 10. IE I have seen many games go 20 moves with one in book, one out. That turns into a two-to-one time handicap. Doubling the time for one program is a _serious_ advantage. Closer to 50-60 Elo based on SSDF numbers.
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.