Author: Russell Reagan
Date: 15:53:32 07/01/03
Go up one level in this thread
On July 01, 2003 at 18:35:55, Omid David Tabibi wrote: >Although I haven't compared the two algorithms, it should be obvious that >verified null-move pruning is more accurate than fail-high reductions, since the >depth reduction is based on a shallow null-move search (dynamic) rather than >static evaluation. More accurate, yes, but "accurate" doesn't mean "stronger" necessarily. If you view the reduced search in FHR as a kind of "verification search", then the main difference between verified null-move and FHR is that verified null-move does a N-R search at first. That should make it more accurate, as you say, but if the static nature of FHR with the reduced verification search is "good enough", FHR will save you the N-R search. So the disadvantage of FHR might be that it doesn't do the N-R search, and the advantage of it might be that it doesn't do the N-R search :) >Verified null-move pruning and fail-high reductions might share some common >ideas (pointed out by Bob Hyatt), but besides that they have little in common >from a practical point of view. You certainly know more about both of these algorithms than I do, but I wonder if there aren't some parts that could be shown equivalent. Maybe there is an improvement upon both of them that could be combined. >Anyway, as ususal the suggestion is: try both of them and use the one that works >better for you. Of course.
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.