Author: Robert Allgeuer
Date: 03:53:44 01/15/04
Go up one level in this thread
On January 15, 2004 at 06:35:45, Tord Romstad wrote: >I have now played a third 100-game match between Gothmog 0.4.5 (which does >not have any learning function) and Phalanx XXII. The results of the three >matches: > >Match on a G4 550 MHz, learning enabled for Phalanx: >Gothmog 0.4.5-Phalanx XXII 52.5-47.5 (+46, -41, =13) > >Match on a PIV 2.4 GHz, learning disabled for Phalanx: >Gothmog 0.4.5-Phalanx XXII 65-35 (+54, -24, =22) > Hi Tord, Are there duplicate games in this test set with learning disabled, meaning that Phalanx and Gothmog have repeated already games played before in the match? >Match on a PIV 2.4 GHz, learning enabled for Phalanx: >Gothmog 0.4.5-Phalanx XXII 51.5-48.5 (+36, -33, =31) > If learning has that big an effect, I would also assume that when learning is enabled the second half of games in the match should be much better for Phalanx than the first 50 games of the match. Did you observe this? Robert >My conclusion: At least against Gothmog's wide and untuned opening book, >Phalanx's learning function is extremely effective, and might be worth >more than 100 Elo points in long matches. > >Tord
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.