Author: Stuart Cracraft
Date: 07:20:10 08/12/04
Go up one level in this thread
On August 12, 2004 at 09:25:11, Anthony Cozzie wrote: >On August 12, 2004 at 07:33:23, Tor Lattimore wrote: > >>Testing with WAC it seems to make very little difference between R=3, R=2 or R=3 >>with verrification, even thought the average ply difference is quite different. >>This leads me to think that R=2 or adaptive based on remaining ply is probobly >>my best bet, but I think i'll do some longer time control tests to see if that >>makes a difference. >>Cheers >>Tor > >Pure R=3 only works for people that do checks in Q-search. > >Examples: Zappa, Gothmog (although tord plays games with his R), and I think >Fruit. > >anthony That is why my GA does and R=3 with verified null move was an improvement over non-verified null move and very nearly equal to plain R=2, as predicted. So it is a permanent setting of my program. Stuart
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.