Author: Andrew Platt
Date: 04:55:58 09/23/04
Go up one level in this thread
On September 23, 2004 at 04:16:35, Tony Werten wrote: >If you allow nullmoves cutoffs during the verification search, you'll still have >problems. Yes, I started to think about this last night. In the original paper by Tabibi and Netanyahu they say to just do a shallower regular search on verify but that uses the same null move (R=3, no material consideration) as the non-verify one. So I'm thinking that I should have two different null move conditions. The first would be when we're not verifying a null move, would use R=3 and wouldn't check material considerations. The second would be when we're verifying a null move that would check material and use R=2; we would use that one when in verify mode. So far verified null move hasn't been a huge saver; the paper does say it comes in more useful at plies 10+ though I haven't seen complete evidence of that. I might just ditch it for now but I'll probably play around with the idea of only using verified null move on later iterations. Andy.
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.