Author: Uri Blass
Date: 05:31:02 07/29/04
Go up one level in this thread
On July 29, 2004 at 08:28:23, Uri Blass wrote: >On July 29, 2004 at 06:41:43, Fabien Letouzey wrote: > >>On July 29, 2004 at 05:58:20, Tord Romstad wrote: >> >>>Consider the extreme case of having an evaluation function which always >>>returns 0. You automatically get perfect move ordering, and you will >>>search really small trees. Your eval will also be very cheap to >>>compute. On the other hand, it could hurt the engine's positional >>>play a bit. :-) >> >>Tord, don't try to confuse the readers. You know move ordering would still >>affect tree size. >> >>Fabien. > >The right words are that you automatically get 100% fail high in the first move >and not that you automatically get perfect move ordering. > >Uri To be more correct 100% of your fail high happen in the first move if you evaluate everything including checkmate as 0. If you fail low in the first move because beta>0 then you fail low on all the other moves. Uri
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.