Author: Uri Blass
Date: 07:23:07 10/06/05
Go up one level in this thread
On October 06, 2005 at 05:26:16, Bruce Moreland wrote: >On October 06, 2005 at 04:10:32, Andreas Guettinger wrote: > >>On October 05, 2005 at 21:03:11, Will Singleton wrote: >> >>>A comp played this move against my comp on ICC today, took less than 30 secs. >>>Can any program find it as fast? >>> >>>[d]3rn1k1/2qbrp1p/1p1ppnpQ/p5NP/1PP1P3/P1N4R/4BPP1/3R2K1 w - - bm Rdd3 >> >>Hiarcs 9.6 UCI plays f4 here. 21s on Mac 400 Mhz. >>I think f4 followed by f5 wins on the spot: f4 Rb8 f5 exf5 Nd5! Qd8 c5! or f4 >>Bc6 f5 exf5 exf5 Qd7 hxg6 fxg6 >> >> >>5 +4.13 bxa5 bxa5 hxg6 fxg6 e5 dxe5 Bd3 (1.68) >> 6 +3.81 bxa5 (3.07) >> 6 +3.67 bxa5 bxa5 hxg6 fxg6 (3.37) >> 6 +3.72 hxg6 (3.43) >> 6 +3.72 hxg6 fxg6 (4.01) >> 6 +3.73 f4 (4.24) >> 6 +3.82 f4 Bc6 Rd4 Rc8 (4.53) >> 6 +3.83 Rhd3 (5.04) >> 6 +3.84 Rhd3 Rc8 Rf3 Bc6 Rf4 d5 (5.36) >> 6 +3.85 Rf3 (6.50) >> 6 +4.09 Rf3 Bc6 hxg6 fxg6 b5 Bb7 Rxf6 Nxf6 (6.70) >> 7 +4.02 Rf3 Bc6 Bd3 axb4 axb4 Ra8 Bc2 (8.27) >> 8 +3.87 Rf3 Bc6 Kh1 d5 exd5 exd5 hxg6 fxg6 Rxf6 Nxf6 (18.75) >> 8 +3.88 f4 (21.80) >> 8 +4.30 f4 Rc8 f5 axb4 axb4 gxf5 exf5 d5 fxe6 fxe6 (29.15) >> 9 +4.32 f4 axb4 axb4 Bc6 f5 exf5 exf5 (59.51) >>10 +4.64 f4 (159.09) >>10 +4.64 f4 Bc6 f5 exf5 exf5 Qd7 fxg6 fxg6 hxg6 (169.23) >>11 +4.89 f4 Rb8 f5 exf5 Nd5 Qd8 Rf1 Bc6 hxg6 fxg6 (541.13) >>12 +4.99 f4 Bc6 f5 exf5 exf5 Qd7 hxg6 fxg6 fxg6 (1570.23) > >If this is the complete line it's not very convincing. If a high score is all >that it takes to earn respect, I'll have mine display +3 on every move. > >The move is probably good, but there is not +5 in the line. > >bruce I think that the position is a bad test position because there are many winning moves. I guess that f4 is one of the moves that leads to forced mate if you search deep enough. You say that there is no +5 in the line. The question is how do you calculate it. A program may consider some positional advantage as even more than +5 and I do not think that it has to be a bad idea. The main question is if it can do mistakes of evaluating +5 positional advantage when the side with the advantage is not winning. If the answer is never or almost never(for example only in less than 0.01% of the cases) then evaluation of more than +5 may be a good idea. For the record Analysis with fruit after f4 with 6 options after 44 minutes at depth 15(analyzing Qc8 at the moment): New game, 3rn1k1/2qbrp1p/1p1ppnpQ/p5NP/1PP1PP2/P1N4R/4B1P1/3R2K1 b - f3 0 1 Analysis by Fruit 2.2: 1. +- (2.02): 1...b5 2.f5 exf5 3.e5 Bc6 4.exf6 Nxf6 5.cxb5 axb4 6.axb4 Be4 7.Ncxe4 fxe4 8.Kh1 d5 9.Rf1 Qb6 10.hxg6 hxg6 2. +- (2.14): 1...axb4 2.axb4 b5 3.c5 Bc6 4.Bg4 Ng7 5.hxg6 fxg6 6.Nxh7 Nxg4 7.Qxg6 dxc5 8.Qxg4 Rxd1+ 9.Nxd1 cxb4 10.Nf6+ Kf7 11.Rh6 3. +- (3.10): 1...Bc6 2.f5 exf5 3.exf5 axb4 4.axb4 Qd7 5.fxg6 fxg6 6.hxg6 Rg7 7.Bd3 Ng4 8.gxh7+ Kh8 9.Qh4 Nef6 10.Re1 4. +- (3.68): 1...Rc8 2.f5 exf5 3.Nd5 Qd8 4.c5 Be6 5.Nxe7+ Qxe7 6.cxd6 Nxd6 7.e5 Nde4 8.exf6 Nxf6 9.hxg6 fxg6 10.Re3 Rc6 11.bxa5 bxa5 5. +- (4.06): 1...a4 2.f5 Bc8 3.hxg6 fxg6 4.Nb5 Qb7 5.fxg6 Rg7 6.gxh7+ Kh8 7.Rf1 Qe7 8.Rhf3 d5 9.Rxf6 Nxf6 10.Rxf6 dxe4 6. +- (3.89): 1...Qc8 2.e5 Bc6 3.exf6 Nxf6 4.Re3 Kh8 5.hxg6 fxg6 6.Qh3 h5 7.bxa5 bxa5 8.Nxe6 Rdd7 9.Qh4 Rf7 (, 06.10.2005) 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.