Author: Russell Reagan
Date: 16:25:17 10/13/03
Go up one level in this thread
On October 13, 2003 at 18:09:36, Steve Maughan wrote: >I think it's only sound if both programs have similar branching factors (i.e. BF >= average ratio of nodes between plys). If they don't then the one with the >smaller branching factors will do better at longer time controls. So I would >imagine that Ruffian v Genius 7 would be much closer at 1 m/g than at 3 min / >move. I overlooked this, but I think this might only be a problem for a small portion of time controls because of the effect of diminishing returns for deeper search depths. For very fast time controls, the branching factor shouldn't be a big factor. For very long time controls, the program with the lower branching factor should reach higher depths, but because of diminishing returns, how valuable are those extra plies at such high depths? There are probably a range of time controls where the effect of diminishing returns is not too great to counteract the effect of the superior branching factor, but at some depth I don't think it will matter. I don't know if this point is a practical time control, and we're getting off into the theoretical never land now, because I doubt we could test this in any reasonable time.
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.