Author: Tom Likens
Date: 09:57:22 01/14/04
Go up one level in this thread
On January 14, 2004 at 12:51:32, Peter Fendrich wrote: >On January 14, 2004 at 07:08:50, Tord Romstad wrote: > >>The day before I released Gothmog 0.4.5, I played a blitz match (4 minutes/game, >>with 1 second increment) between my engine and Phalanx XXII on my PowerBook >>G4 550 MHz. Gothmog narrowly won the match; the final score was 52.5-47.5. >> >>Yesterday, I started a new match between the same two engines. The only >>difference was that this match was played on a PIV 2.4 GHz, and that learning >>was disabled for Phalanx in the second match (my own engine doesn't learn). >>The result: 65-35 for Gothmog. >> >>Is this just a statistical fluctuation, or is learning really that effective? >>Or perhaps Phalanx (a very old engine) simply doesn't play well on fast >>hardware? >> >>Tord > >Of course it can be random fluctuations but i have run several matches with and >without learning. I can tell you that aggressive learning vs an opponent with a >tiny book will eventually crush that opponent. The same goes if the opponent >have a bad book. It makes a big difference quickly. Aggressive learning vs an >opponent with a good big wide and "bushy" book is not at all that effective. It >can even be the other way around you will "overlearn" unimportant things. Less >aggressive learning is better in that case and it will then take a longer time >before you see any results from the learning. >I have no idea what kind of learning Phalanx is using. >/Peter How odd Peter, I hadn't realized that you had responded and wrote about my recent experience with Terra. My compliments on your learning algorithm, it's very effective (or maybe Djinn uses the quintessential bad book you mention above) ;-) regards, --tom
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.