Author: Jeremiah Penery
Date: 05:03:41 05/21/00
Go up one level in this thread
On May 20, 2000 at 14:52:56, Enrique Irazoqui wrote: >On May 20, 2000 at 14:08:51, Ricardo Gibert wrote: > >>On May 20, 2000 at 13:20:47, Bruce Moreland wrote: >> >>>On May 20, 2000 at 10:04:58, Ricardo Gibert wrote: >>> >>>>On May 20, 2000 at 09:27:39, Robert Hyatt wrote: >>>> >>>>>On May 20, 2000 at 07:11:31, Terje Vagle wrote: >>>>> >>>>>>1r6/5kp1/RqQb1p1p/1p1PpP2/1Pp1B3/2P4P/6P1/5K2 b - - >>>>>> >>>>>>After 45. Ra6, Fritz suggests Qe3 for black and evaluates the position as 0,94. >>>>>>It does not seem to find the famous draw-line for Kasparov. >>>>>>10 hours analysis on PIII-600, and 28006383 KN evaluated >>>>>> >>>>>>Does any other program find the draw-line? >>>>>> >>>>>> >>>>>>Regards >>>>>> >>>>>> >>>>>>Terje >>>>> >>>>>The draw is somewhere around 60 plies deep, total. I don't think anyone is >>>>>going to find that. >>>> >>>>60 plies? No way. More like half that. A lot of humans have no trouble >>>>calculating the draw to the end. I found it myself without any assistance of any >>>>kind. The only thing difficult about that draw is the psychological barrier. >>>>This Kaspy failed to overcome as we all know. >>>> >>>>As for computer programs, judging from the Chris Janeke post, commercial >>>>programs have no trouble finding the draw with a 5 ply headstart. If it were >>>>really 60 plies, I don't see how that would be possible. >>> >>>Oh yeah? You'd see the part that goes h4 h5? It's not that easy. I helped >>>analyze this with a GM and an IM the day it happened, and they spent hours >>>trying to prove that there was a win. >> >>Yeah. I found h4 too and I remember the only reply I considered was h5. > >Is this the drawing line? The evaluation of Fritz 6b is not 0.00, but it's >getting there. > >Enrique > >Deep Blue - Kasparov,G >1r6/5kp1/RqQb1p1p/1p1PpP2/1Pp1B3/2P4P/6P1/5K2 b - - 0 1 > >Analysis by Fritz 6: > >45...Qxc6-- > ± (0.72) Depth: 1/3 00:00:00 >45...Qxc6-- > ± (0.72) Depth: 1/6 00:00:00 >45...Qxc6 46.dxc6 Bc7 > ² (0.47) Depth: 2/7 00:00:00 >45...Qxc6 46.dxc6 Bc7 47.Ra7 > ² (0.62) Depth: 3/10 00:00:00 >45...Qxc6 46.dxc6 Bc7 47.Ra7 Rc8 > ² (0.59) Depth: 4/10 00:00:00 1kN >45...Qxc6 46.dxc6 Bc7 47.Ra7 Rc8 48.g3 > ² (0.69) Depth: 5/12 00:00:00 3kN >45...Qxc6-- > ± (1.00) Depth: 6/16 00:00:00 8kN >45...Qxc6-- 46.dxc6 Be7 47.Bd5+ Ke8 48.c7 Rc8 49.Ra7 > ± (1.12) Depth: 6/16 00:00:00 11kN >45...Qxc6 46.dxc6 Bc7 47.Ra7 Rc8 48.Rb7 Ke7 49.Rxb5 > ± (1.22) Depth: 7/18 00:00:00 31kN >45...Qxc6 46.dxc6 Bc7 47.Ra7 Rc8 48.Rb7 Ke7 49.Ke2 Kd6 50.Rxb5 > ± (1.28) Depth: 8/16 00:00:00 59kN >45...Qxc6 46.dxc6 Rd8 47.Ra7+ Kf8 48.Rb7 Rb8 49.Rd7 > ± (1.37) Depth: 9/18 00:00:00 131kN >45...Qxc6 46.dxc6 Kf8 47.Ra7 Rc8 48.Rb7 Ra8 > +- (1.44) Depth: 10/21 00:00:00 309kN >45...Qxc6 46.dxc6 Kf8 47.Ra7 Rc8 48.Rb7 Ra8 > +- (1.44) Depth: 11/22 00:00:01 517kN >45...Qxc6 46.dxc6 Kf8 47.Ra7 Rc8 48.Rb7 h5 49.Rxb5 Ke7 50.Bd5 > +- (1.62) Depth: 12/24 00:00:05 2353kN >45...Qxc6 46.dxc6 Kf8 47.Ra7 Rc8 48.Rb7 Ke8 49.Rxg7 Ra8 50.c7 > +- (1.72) Depth: 13/27 00:00:08 4066kN >45...Qxc6 46.dxc6 Kf8 47.Ra7 Rc8 48.Rb7 h5 49.Rxb5 Ke7 > +- (1.72) Depth: 14/26 00:00:14 7078kN >45...Qxc6 46.dxc6 Kf8 47.Ra7 Rc8 48.Rb7 h5 49.Rxb5 Ke7 50.Ra5 > +- (1.75) Depth: 15/28 00:00:26 12926kN >45...Qe3! > +- (1.72) Depth: 15/41 00:01:33 43976kN >45...Qe3! 46.Qd7+ Kg8 47.Qxd6 Rf8 48.Qe6+ Kh7 49.Bf3 Qc1+ 50.Kf2 > ± (1.25) Depth: 15/41 00:02:03 57808kN >45...Qe3 46.Qxd6 Re8 47.h4 h5 48.Bf3 Qc1+ 49.Kf2 Qd2+ 50.Be2 > +- (1.53) Depth: 16/42 00:05:13 146128kN >45...Qe3 46.Qxd6 Re8 47.h4 h5 48.Bf3 Qc1+ 49.Kf2 Qd2+ 50.Be2 > ± (1.28) Depth: 17/46 00:13:16 369407kN >45...Qe3 46.Qxd6 Re8 47.h4 h5 48.Bf3 Qc1+ 49.Kf2 Qd2+ 50.Be2 > ± (1.16) Depth: 18/46 00:32:34 906397kN > >(P600E/202MB, Cadaqués 20.05.2000) Well, this is what Ed gives on the Rebel pages: "Analysing this with Rebel we tried to find out the longest defense till the draw (repetition) would be found. Well it is at least 36 plies. Even for DB super machine that's too much. The line goes like this: 45.Ra6? Qe3! 46.Qxd6 Re8! 47.h4! h5! 48.Bf3 Qc1+ 49.Kf2 Qd2+ 50.Be2 Qf4+ 51.Kg1 Qe3+ 52.Kh2 Qf4+ 53.Kh3 Qxf5+ 54.Kh2 Qf4+ 55.Kg1 Qe3+ 56.Kf1 Qc1+ 57.Kf2 Qf4+ 58.Ke1 Qc1+ 59.Bd1 Qxc3+ 60.Kf1 Qc1! 61.Ke2 Qb2+ 62.Kf1 Qc1 Note 60..Qc1! It's a (quiet) non checking move. Perhaps the reason why no program can find 45..Qe3 with a draw score?" With some version of Crafty a while back, I let it run for a week or two on my machine to see if it could find the draw. It followed Ed's line for a while (up to ply 20 or so), and the evaluation actually dropped to 0.00 at one point. Unfortunately, it could not see 60. ...Qc1!! and thought it could find a way out of the perpetual. The score went back up at about ply 21, and continued to rise through about 23, where I stopped it. It was back to +1.xx. I don't think any program is capable of finding the entire correct line here. Just as Ed says, I think it's because of the non-checking move 60. ...Qc1!! that gets pruned this deeply in the tree. Jeremiah
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.