Author: Vincent Diepeveen
Date: 06:09:27 06/17/01
Go up one level in this thread
On June 17, 2001 at 04:11:26, Uri Blass wrote: >Are there top programs that are designed to see always simple tactics after >enough time? > >Is it possible to change them to see always simple tactics after enough time >when the price is only being 0.1% slower by using 2 engines when one engine is >usually used only for only 0.1% of the time or maybe my math is wrong and only >using another engine makes the program already more than 0.1% slower because of >the fact that the program is bigger(even if the other engine is simply the >original engine without null move pruning)? Here i can jump in with double nullmove. A few years ago some people doubted nullmove to be a correct way of searching in *any* respect. Nothing would beat fullwidth. Hence Deep Blue being fullwidth. Nowadays thinking has changed after everyone has been beaten by nullmove. Default nullmove implementation of course will have major problems seeing the consequences in your position, as it can't detect zugzwang and suffers from a reduction factor (the latter being the smallest problem). I do use nullmove in endgames except pawn endgame (where there are too many zugzwangs and where you search deep anyway). Here diep with 150mb hash 2 processors and no EGTBs. diep without egtb: Proces 1 starting initialisation! Proces 1 starting SearchSMP! Took 0.08 seconds to start all 1 other processes Process 1: engineflags=0 msk=1 process 0: engineflags = 0 denktime=10000000 maxtime=10000000 00:00 0 0 3 (0) 1 2.817 Nb8-c6 00:00 0 0 7 (0) 1 3.122 Ke4-d4 00:00 0 0 70 (0) 2 2.701 Ke4-d4 Ne1-f3 Kd4xc5 00:00 0 0 181 (0) 3 2.701 Ke4-d4 Ne1-f3 Kd4xc5 ++ b8-c6 00:00 0 0 279 (0) 3 3.517 Nb8-c6 Kg2-h3 f4-f3 00:00 0 0 743 (3) 4 3.517 Nb8-c6 Kg2-h3 f4-f3 Kh3xh4 00:00 0 0 2955 (330) 5 5.351 Nb8-c6 Kg2-h3 f4-f3 Ne1-c2 g3-g2 Kh3-h2 Ke4-d3 00:00 0 0 15100 (3402) 6 6.303 Nb8-c6 Kg2-h1 h4-h3 Ne1-g2 f4-f3 Ng2-h4 00:00 0 0 50553 (10477) 7 7.130 Nb8-c6 Kg2-h3 f4-f3 Ne1-c2 f3-f2 Kh3-g2 h4-h3 Kg 2xg3 f2-f1Q b7-b8Q 00:01 0 0 140198 (20614) 8 7.746 Nb8-c6 Kg2-f1 h4-h3 Kf1-g1 f4-f3 Ne1xf3 Ke4xf3 Kg1-h1 h3-h2 00:03 0 0 358299 (57862) 9 7.798 Nb8-c6 Kg2-g1 h4-h3 Kg1-f1 f4-f3 Ne1xf3 Ke4xf3 Kf1-g1 h3-h2 Kg1-h1 00:22 0 0 2361968 (712131) 10 10.424 Nb8-c6 b7-b8Q Nc6xb8 Kg2-h1 h4-h3 Ne1-c2 f4 -f3 Nc2-a3 Ke4-d4 Na3-c2 Kd4xc5 Nc2-e3 g3-g2 Kh1-g1 00:25 0 0 2698143 (724893) 11 11.406 Nb8-c6 b7-b8Q Nc6xb8 Kg2-h3 f4-f3 Ne1-g2 f3 -f2 Kh3xh4 f2-f1Q Kh4xg3 Nb8-c6 Kg3-h2 00:32 0 0 3537816 (739317) 12 12.901 Nb8-c6 b7-b8Q Nc6xb8 Kg2-g1 h4-h3 c5-c6 Nb8 xc6 Ne1-c2 h3-h2 Kg1-h1 f4-f3 Nc2-e1 Ke4-e3 Ne1-c2 Ke3-f2 00:48 0 0 5627926 (1254899) 13 13.380 Nb8-c6 b7-b8Q Nc6xb8 Kg2-h3 f4-f3 Ne1-g2 f 3-f2 Kh3xh4 f2-f1Q Kh4xg3 Nb8-c6 Ng2-h4 Qf1-e2 Kg3-h3 Qe2-f2 Nh4-g2 01:24 0 0 9896003 (1503623) 14 14.713 Nb8-c6 b7-b8Q Nc6xb8 Kg2-h3 f4-f3 Ne1-g2 f 3-f2 b2-b4 f2-f1Q Kh3-g4 >The position in the end of my post can happen in a game(it happened in an >engine-engine game when I started from one of the position from my >correspondence games). > >Deep Fritz could not see simple tactics because of null move pruning(with >selectivity=0 Deep fritz has no problem to see more than 10 pawns advantage for >black by Nc6 but with null move pruning Nc6 failed low again and again and fritz >change it's mind to Ke3 when it is not clear if black can win after this move). > >I remember that Amir Ban said about other rare cases when Junior cannot see >simple tactics that it is not a bug and I understand that he is not going to fix >it so I wonder if there is a top program that is designed to see always simple >tactics after enough time. > > >New position >[D]1n6/1P6/8/2P5/p3kp1p/6p1/1P4K1/4N3 b - - 0 1 > >Analysis by Deep Fritz: > >1...a3-- > ³ (-0.31) Depth: 1/3 00:00:00 >1...a3-- > ³ (-0.31) Depth: 1/3 00:00:00 >1...f3+! > ³ (-0.38) Depth: 1/3 00:00:00 >1...h3+! > ³ (-0.50) Depth: 1/3 00:00:00 >1...h3+! 2.Kxh3 > µ (-0.91) Depth: 1/3 00:00:00 >1...Nc6! > -+ (-1.69) Depth: 1/3 00:00:00 >1...Nc6! > -+ (-2.13) Depth: 1/3 00:00:00 >1...Nc6-- > -+ (-1.81) Depth: 2/4 00:00:00 >1...Nc6-- > -+ (-1.81) Depth: 2/4 00:00:00 >1...Kd4! > -+ (-2.09) Depth: 2/6 00:00:00 >1...Kd4! 2.Nf3+ > -+ (-2.22) Depth: 2/6 00:00:00 >1...Kd4 2.Nf3+ Kxc5 3.Nxh4 > -+ (-2.22) Depth: 3/8 00:00:00 >1...Kd4 2.Nf3+ Kxc5 3.Nxh4 Kd5 > -+ (-2.22) Depth: 4/10 00:00:00 >1...Kd4 2.Nf3+ Kxc5 3.Nxh4 Kc6 4.Ng6 > -+ (-1.97) Depth: 5/11 00:00:00 >1...Kd4-- > -+ (-1.66) Depth: 6/12 00:00:00 2kN >1...Kd4-- 2.Nf3+ Kxc5 3.Nxh4 Kc6 4.Ng6 Kxb7 > -+ (-1.47) Depth: 6/13 00:00:00 2kN >1...Kd5! > -+ (-1.50) Depth: 6/13 00:00:00 2kN >1...Kd5! 2.Nd3 Kd4 3.Nxf4 Kxc5 4.Ne6+ Kd6 > -+ (-1.75) Depth: 6/13 00:00:00 3kN >1...Nc6! > -+ (-1.78) Depth: 6/13 00:00:00 4kN >1...Nc6! 2.Kh3 f3 3.Nxf3 > -+ (-2.13) Depth: 6/13 00:00:00 6kN >1...Nc6 2.Kh3 f3 3.Nxf3 Kxf3 4.Kxh4 g2 > -+ (-2.41) Depth: 7/15 00:00:00 8kN >1...Nc6-- > -+ (-2.09) Depth: 8/10 00:00:00 10kN >1...Nc6-- > -+ (-2.09) Depth: 8/18 00:00:00 12kN >1...Nc6! > -+ (-2.41) Depth: 9/17 00:00:00 28kN >1...Nc6-- > -+ (-2.09) Depth: 10/14 00:00:00 37kN >1...Nc6-- > -+ (-2.09) Depth: 10/19 00:00:00 45kN >1...Nc6! > -+ (-2.41) Depth: 11/20 00:00:00 87kN >1...Nc6-- > -+ (-2.09) Depth: 12/21 00:00:00 128kN >1...Nc6-- > -+ (-2.09) Depth: 12/24 00:00:00 172kN >1...Nc6-- > -+ (-1.78) Depth: 13/21 00:00:00 261kN >1...Nc6-- > -+ (-1.78) Depth: 13/26 00:00:00 377kN >1...Nc6-- > -+ (-1.47) Depth: 14/24 00:00:01 660kN >1...Nc6-- > -+ (-1.47) Depth: 14/28 00:00:01 922kN >1...Nc6-- > µ (-1.16) Depth: 15/27 00:00:02 1600kN >1...Nc6-- > µ (-1.16) Depth: 15/30 00:00:03 2180kN >1...Nc6-- > µ (-0.84) Depth: 16/29 00:00:05 3386kN >1...Nc6-- > µ (-0.84) Depth: 16/30 00:00:06 4622kN >1...Ke3! > µ (-0.88) Depth: 16/35 00:00:11 7839kN >1...Ke3! 2.Nc2+ Kd3 3.c6 Nxc6 4.Nb4+ Ke2 5.Nxc6 f3+ 6.Kh3 f2 7.b8Q > µ (-0.97) Depth: 16/39 00:00:13 9840kN, tb=1 >1...Ke3 2.Nc2+ Kd3 3.c6 Nxc6 4.Nb4+ Ke2 5.Nxc6 f3+ 6.Kh3 f2 7.b8Q > µ (-0.88) Depth: 17/37 00:00:23 17184kN, tb=4 > >(blass, telaviv 16.06.2001) > >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.