Author: Ernst A. Heinz
Date: 04:12:32 08/03/98
Go up one level in this thread
On August 03, 1998 at 06:48:06, Robert Hyatt wrote: >On August 03, 1998 at 03:14:03, Ernst A. Heinz wrote: > >>On August 03, 1998 at 02:21:12, Amir Ban wrote: >>> >>>I always accept the fail-high move immediately. >>> >>>There are parameters that you are important that you don't mention: What size >>>window is used during normal search ? If it's zero-width, or very small, then >>>fail-high doesn't mean it's much better than the previous best, and you can take >>>the new move or leave it. If you use a 0.3-0.5 window, as I do, fail-high means >>>it's clearly better than the previous, even it later fails low. >>> >>>Also, what window do you use on the re-search that fails low ? If you use >>>new-alpha+1 to infinity, then maybe the fail-low happened because the value is >>>exactly new-alpha. If you use a window of old-alpha to infinity, then a >fail-low >>>indeed makes the move suspicious. >>> >>>Amir >> >>Good point, Amir! >> >>Actually, I do not recall to having encountered the fail-low after fail-high >>behaviour as long as the research is done with the *old alpha*, i.e. the >>same that was used when the fail-high occured. >> >>Otherwise, this nasty fail-low after fail-high behaviour unfortunately >>seems to occur quite frequently. >> >>=Ernst= > >This is a known problem will null-movers... if you use the old alpha, you may >not get the fail low, but you will get a score lower than the previous best >move, which might be just as ugly... No, I meant that with our current "DarkThought" I do not recall having seen the top-level research drop below the score of the previously best move + 1 as long as it was done with the *old* alpha. =Ernst=
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.