Author: Robert Hyatt
Date: 15:47:24 01/22/98
Go up one level in this thread
On January 22, 1998 at 14:08:34, Bruce Moreland wrote: > >On January 22, 1998 at 14:00:36, Robert Hyatt wrote: > >>The math part is simple. The PV takes longer to search than all the >>rest of the moves, in general. If you have 2 PV's, the math is a >>killer. >>Would be real easy to test, just skip the first move at the front of >>each >>iteration and then come back at the end and pick it up. > >He is searching the successors with a null window, Bob. > doesn't help... notice that a fail high still takes N times as long as a normal fail low. And *some* of those root moves may well fail high. 1. what initial window can be used? Old score from last iteration? Can be off significantly with an extra ply added on. If the initial window is too low, fail highs will be killingly costly. If it is too low, everything might fail low and there's no useful info. If the supposedly best move (last one searched) returns a value lower than expected, what do you do about all the others that failed low on that null window? 2. Suppose all moves fail low, including the supposed best one. research them all? IE I don't know how to make this efficient at all. Alpha/beta depends on alpha and beta to limit the size of the tree. This approach seems to need a hearty dose of witchcraft as well... :) >burce
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.