Author: Serge Desmarais
Date: 15:21:33 08/27/98
Go up one level in this thread
On August 24, 1998 at 01:45:32, Jouni Uski wrote: >In weekend I run whole 231 position suite with 3 minute limit in my P90 with >three programs. Here are results and comparison to 1 minute run: > > 1 min 3 min >Fritz5.00 116 145 (+29) >Nimzo98(Pad.) 100 131 (+31) >Mchess 7 89 118 (+29) > >There is one dubious thing, when running testsuite automatically within Fritz. >Even if You define extra ply = 4, sometimes immediately after founding move >solution time is recorded and search stopped. So You cannot 100% trust Fritz >results. Why can't Fritz simple search until time limit? > >Jouni I noticed that with most Fritz programs, when they see a mate, they STOP calculating and play the move, even if there is a quicker mate to find. But on its next move (in a game) its recalculates everything... Opposed to that, Genius keeps searching and does find the faster mate. After the mate is found, it plays all the moves instantly to the mate, while Fritz has to recalculate the mate at every move. You apparently cannot make Fritz contimue calculating after it found a forced mate. Well there is maybe a way, in infinite analysis, if you add up new variations to be displayed in the window, it continues calculating for these and as it goes deeper, the first line (the one or those with the forced mate) are also updated and so it will find the fastest mate, eventually... Serge Desmarais Serge Desmarais
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.