Author: Ed Schröder
Date: 11:01:21 12/30/02
Go up one level in this thread
On December 30, 2002 at 11:32:08, Vincent Diepeveen wrote: >On December 29, 2002 at 19:35:52, Martin Giepmans wrote: > >If you use nullmove don't use any of the reduction-2 stuff. >Ed isn't using nullmove. It is very dangerous to use. :-) I gave a warning, from the page: OTHER REDUCTIONS : REBEL does several other reductions, to understand them you will need to know a little bit more about the quite different approach of REBEL's Search Algorithm in comparison to other chess programs, I therefore highly question the relevance of this topic as I assume that it can't be used in other programs without drastic changes, nevertheless here goes... ========== But wait till I have explained null-move, you will have a better picture then. Ed >I also concluded the red1 doesn't work too well for me. > >I get 1-5% speedup out of it. > >It basically only reduces the tree after for example from opening > >e4 h5 qxh5 rxh5 d4 <reduction-1> <nullmove> (some moves and cutoff) > >So my nullmove reduction factor becomes from R=3 to R=4 in such a case. > >If i want to use R=4 i'll let my program know it by myself, >so i turned off reduction-1 for now. It's a feature however in diep people >can turn on when they want to. > >>On December 29, 2002 at 19:05:51, Tony Werten wrote: >> >>>http://members.home.nl/matador/chess840.htm >>> >>>Tony >> >>The formula for reductions 2b and 2c look strange: >>"alpha < score + threat + margin --> reduce depth ..." >> >>This would imply that a bigger threat gives more reductions ... >>Should be score - threat - margin? >> >>Martin
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.