Author: Bas Hamstra
Date: 16:40:34 05/18/01
Go up one level in this thread
On May 18, 2001 at 18:39:46, Dann Corbit wrote: >On May 18, 2001 at 18:26:08, Bas Hamstra wrote: > >>After 3 rounds Fritz leads with 3 out of 3. CT and Diep follow with 2.5 out of >>3. Nice work Vincent! >> >>To my pleasant suprise Tao won from Goliath (0.2M nps against an incredible 2M >>nps from Goliath). Tao lost then from Fritz (after 6 billion moves). > >That will certainly update my database for "longest game" >;-) :-) The game was kind of stuck with Fritz in a better position, but it was difficult to make progress. After a series of non-moves, the game was decided in the last half hour when Fritz finally got a decisive attack against the weakened black king. BTW I must say Vincent predicted the result already at move 2, explaining that Black was strategically lost. Ok, maybe it was move 12, I'm not quite sure :-) However *HAD* white played dxe6, he pointed out, it would have been the other way around, and White would have been absolutely and totally lost (strategically speaking). Since the sum of the ratings of us three was about the half of Vincent's we humbly bowed our ignorant heads. "But isn't this freepawn at least worth 0.001 pawn?" I dared to ask. But no, it appeared that the freepawn was the mother of all blocks of concrete, hanging on the leg of the position and trying to swim was ridiculous. >I think so far Tao and Quark are the real stunners, and playing very amusing >chess. Hearty congratulations to all of the participants for all the hard >work. Thanks. I am trying to teach it to attack, inpired by Comet's shameless king attacks. For Comet a game simply consists of opening, king attack and endgame. Diep does a good job too. The games they play are far more fun to watch than the average bloodless computer game. I hope I can move my program somewhat in that direction, even if it costs me some rating points. But I have noticed this. If you put some extra knowledge in, that takes (say) 10% extra evaluation time, it may occur that you suddenly search dramatically less deep. Suddenly it sees all kinds of complications. If you don't pay attention the resulting program may play worse because it gets outsearched. I wonder how Fritz and Comet deal with this. Diep has already so much eval that it doesn't matter anyway. Tomorrow morning I play YACE! Best regards, Bas.
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.