Author: pavel
Date: 15:52:18 04/15/05
Go up one level in this thread
On April 15, 2005 at 18:43:50, Uri Blass wrote: > >What is your target? > Both is my target. Because if I play 100,000s of games than I will be closer to the 1 elo margin at the same time I will see the effect of positional learning in the long term. My primary target is to test learning effect, because from what I understand it will take atleast 500,000 or more to reach a 1 elo margin; which probably will be too much of a stress. Do you have plans to implement positional learning in Movei? Cheers, pavs >If your target is to test learning then I understand matches of 100000 games but >if your target is to achieve 1 elo margin then you achieve nothing because >an engine that is weaker in tournament against other engines can beat stronger >engine thanks to better learning. > >I do not think that 30,000 is a low limit because almost nobody play more than >30,000 games. > >even if the average time of the game is 10 minutes then you still need many >months for 30,000 games and after many months there are already new versions and >the old versions are not interesting. > >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.