Author: Bas Hamstra
Date: 02:14:44 02/15/00
Go up one level in this thread
On February 14, 2000 at 13:03:27, Will Singleton wrote: >On February 14, 2000 at 10:35:48, Bas Hamstra wrote: > >>On February 13, 2000 at 03:35:50, Will Singleton wrote: >> >>>I am always interested in the nps kibitzed by programs playing on ICC. You can >>>never be sure as to the accuracy, since one can fiddle with the numbers. But I >>>was surprised the other day to see an amateur program getting in excess of 1 >>>million nps on a 700mhz machine. Many amateur progs get around 100knps or less. >>> >>>Just wondering. Anybody get close to or more than 1000 nps per mhz? >>> >>>Will >> >>I get 400k on a Celeron 466 with mostly material/PieceSquare. With decent eval I >>expect to do around 250k. >> >>BTW does everyone count nodes the same way? I for example don't doublecount D==0 >>nodes. Therefore, before *every* qsearch I do a Nodes-- to eliminate the >>doublecount. My NPS would go way up if I did't do this decrement. >> >> >>Regards, >>Bas Hamstra. > >Haven't heard from you lately, was wondering where you were with Tao. I have done some experiments with a completely stripped down eval. Not nice to watch :) Basically only a piece square table. Still it managed a 2190 FICS rating at Blitz. The piece square table is such, that the program tries to advance its pieces in enemy direction as far as possible. Expecially knights, of course, but also all other pieces. What happens if a program thinks *very* deep and *only* tries to put pieces (and cde pawns) on your half of the board? Well it does not too bad, even against humans. It plays sometimes ugly moves. It gets a reasonable blitz rating. It gets whacked fast by some programs with more knowledge (The Crazy Bishop!) even if they are much slower. On the whole, I must say it did far better than I expected. Just an experiment of course, I see it rather play sound and aggressive *real* chess :) Regards, Bas Hamstra. >I assume that you increment nodes on each call to search, thereby necessitating >the decrement before going to qsearch. I just count nodes in make_move, so I >don't have to deal with that. > >Will
This page took 0.11 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.