Author: Vine Smith
Date: 06:20:19 04/19/02
Go up one level in this thread
On April 19, 2002 at 05:09:15, Vine Smith wrote: >On April 19, 2002 at 04:18:07, Uri Blass wrote: > >>On April 19, 2002 at 03:46:59, Sune Larsson wrote: >> >>>On April 19, 2002 at 03:39:29, Otello Gnaramori wrote: >>> >>>>On April 18, 2002 at 20:31:24, Vine Smith wrote: >>>> >>>>>If it weren't for the truncated time control, Smirin >>>>>could be leading 3-0. >>> >>> >>>> >>>>This is IMHO sincerely exaggerated, since also at regular timing e.g. GM Huebner >>>>was drawish in all the games against Deep Fritz in the last match. >>>> >>>>w.b.r. >>>>Otello >>> >>> >>> "Sincerely exaggerated"?? Did you look at these 3 games played between >>> Smirin vs programs? If yes, do you understand these games? IMHO Vine is >>> totally correct here. >>> >>> Sune >> >>It is not clear because it is possible that the computers could also play better >>with more time. >> >>Uri > >Naturally, it's all conjecture, but it does seem fairly clear that longer time >controls favor humans more than programs. To compare extremes, while you've >shown that programs can compete with correspondence GMs, I think the programs >have little chance of racking up the kinds of ratings there that they achieve by >blitzing at ICC and FICS (even excluding the ones that are artificially >inflated). In an interview prior to the match (on the Kasparovchess site, as I >recall), Smirin sounded displeased with the short time control he had inherited >from Gulko. >And when you look at what Smirin has been doing, it appears he has been quite >successful at turning the games into wars of ideas rather than tactics. Shredder >would not have understood what was happening to it given much more time, I >believe, although it may have sensed the danger a little earlier, still in a >rather hopeless position. Similarly, Hiarcs 8 (about which we know very little >prior to its release) did not seem to appreciate the weakness of its advanced >d-pawn, which again is a long-term danger that probably would not trigger alarms >in a longer search until too late. Junior possibly would not have misplaced its >queen given more time to think -- I think I will check this out. Without the >threats of trapping the queen, Smirin's kingside pressure would have been >somewhat easier to deal with, but he still would have held the initiative. >In any case, there's nothing drawish about the kind of chess Smirin has been >playing, as asserted toward the start of this thread. I think he's quite >interested in filling his pockets with as much of Chessbase's cash as possible, >and is going about this very cleverly. > >Regards, >Vine This was run on a PIII-850 using 128 Mb hash, and so is much slower than the system being used for the Smirin match. I let it run for around half an hour each time. Of course, it's not possible to duplicate whatever was floating around in the hash during the game either. That being said: Ilya Smirin - Deep Junior r1r3k1/1p1qbpp1/2n2n1p/pNp5/2PpPP2/PP1P1N2/6KP/R1BQ1R2 b - - 0 1 Analysis by Junior 7: 24...h5 25.f5 = (-0.24) Depth: 3 00:00:00 24...b6 25.f5 ³ (-0.30) Depth: 3 00:00:00 24...a4 25.f5 axb3 ³ (-0.40) Depth: 3 00:00:00 24...Qg4+ 25.Kh1 ³ (-0.45) Depth: 3 00:00:00 24...Qg4+ 25.Kh1 ³ (-0.45) Depth: 6 00:00:00 24...Qg4+ 25.Kh1 ³ (-0.45) Depth: 9 00:00:00 24...Qg4+ 25.Kh1 ³ (-0.45) Depth: 12 00:00:00 5kN 24...Qg4+ 25.Kh1 Qh5 26.Rg1 = (-0.18) Depth: 15 00:00:08 4622kN 24...g6 25.h3 Kh7 26.Qd2 Bf8 27.e5 Nh5 28.Rg1 Be7 = (-0.19) Depth: 15 00:00:16 9807kN 24...g6 25.h3 a4 26.bxa4 Qe6 27.Bd2 Kh7 28.Rb1 Re8 = (-0.16) Depth: 17 00:00:38 23580kN 24...g6 25.h3 a4 26.bxa4 Ne8 27.Bd2 Kh7 28.Rb1 Nf6 29.f5 gxf5 = (-0.12) Depth: 18 00:02:29 86394kN 24...g6 25.h3 a4 26.bxa4 Kh7 27.Rb1 Re8 28.Ne5 Qc8 29.Nxf7 Qe6 30.Ne5 Nxe5 31.fxe5 = (-0.09) Depth: 19 00:07:08 243562kN 24...g6 25.h3 Nh7 26.f5 gxf5 27.exf5 Kh8 28.Kh1 Rg8 29.Bxh6 Rg3 = (-0.01) Depth: 20 00:16:34 555424kN 24...g6 25.h3 Nh7 26.Bd2 h5 27.Rg1 Kh8 28.a4 Re8 29.Qe2 h4 = (-0.02) Depth: 21 00:39:07 1314126kN (Smith, Stevenson Ranch 19.04.2002) 24...Qg4+ wasn't criticized in the annotations at Chessbase, but somebody needs to explain why this check is not bad. Was Smirin's king on g2 so strongly posted that it needed to be driven away? Looking at the position, and the game continuation, isn't it obvious that Smirin intends Kh1 in any event, so this is just a lost tempo? Well, it's not obvious to Junior, which does change its mind in 8 seconds, but I'm not sure that 24...g6 is really an improvement, since it provides a contact point for an f5 advance, as well as weakening h6. But most telling is the reply for White being predicted in all plies -- 25.h3, as if the check still needed to be stopped, even at the cost exposing White's king more, rather than less. Ilya Smirin - Deep Junior r1r3k1/1p2bpp1/2n2n1p/pNp5/2PpPPq1/PP1P1N2/7P/R1BQ1R1K b - - 0 1 Analysis by Junior 7: 25...h5 26.f5 ³ (-0.63) Depth: 3 00:00:00 25...b6 26.f5 ³ (-0.69) Depth: 3 00:00:00 25...a4 26.f5 axb3 µ (-0.77) Depth: 3 00:00:00 25...Qh3 26.f5 µ (-1.00) Depth: 3 00:00:00 25...Qh3 26.Rg1 = (-0.02) Depth: 6 00:00:00 1kN 25...Qd7 26.f5 Re8 = (-0.08) Depth: 6 00:00:00 2kN 25...Qd7 26.Kg2 = (0.00) Depth: 9 00:00:00 8kN 25...Qh3 26.Rg1 = (-0.02) Depth: 9 00:00:00 10kN 25...Qh3 26.Rg1 = (-0.02) Depth: 12 00:00:00 79kN 25...Qh3 26.Rg1 = (-0.02) Depth: 15 00:00:01 994kN 25...Qh3 26.Rg1 = (-0.02) Depth: 18 00:00:20 12443kN 25...Qh3 26.Rg1 Nh5 27.Rg2 a4 28.Ng1 Qh4 29.Nf3 = (0.00) Depth: 20 00:03:13 111990kN 25...Qh3 26.Ra2 Ng4 27.Rg2 Ne3 28.Bxe3 dxe3 29.Ng1 Qh4 30.Qf3 a4 31.bxa4 Rd8 ² (0.28) Depth: 21 00:27:45 913467kN (Smith, Stevenson Ranch 19.04.2002) Here Junior shows evidence it might have retreated the queen if only searching at a very shallow depth, but then fixes on 25...Qh3, a square where it eventually got into trouble, at plies 9 and above. Possibly an effect of high king tropism, meant to encourage aggressive play by Junior? Ilya Smirin - Deep Junior r1r3k1/1p2bpp1/5n1p/pPp5/2PpPP2/1P1P1N1q/R6P/2BQ1R1K b - - 0 1 Analysis by Junior 7: 28...h5 29.f5 ³ (-0.64) Depth: 3 00:00:00 28...b6 29.f5 ³ (-0.70) Depth: 3 00:00:00 28...Ng4 29.f5 µ (-0.71) Depth: 3 00:00:00 28...Re8 29.f5 µ (-0.77) Depth: 3 00:00:00 28...Re8 29.Rg2 ² (0.26) Depth: 6 00:00:00 28...Ng4 29.Ng1 = (-0.03) Depth: 6 00:00:00 28...h5 29.f5 = (-0.04) Depth: 6 00:00:00 1kN 28...Rf8 29.Ng1 Qg4 30.Qxg4 Nxg4 ³ (-0.29) Depth: 6 00:00:00 2kN 28...Rf8 29.Rg2 ² (0.42) Depth: 9 00:00:00 7kN 28...h5 29.f5 Ng4 = (-0.03) Depth: 9 00:00:00 13kN 28...h5 29.Rg1 Re8 30.Rg3 Qc8 31.f5 Bd6 32.Rgg2 = (0.25) Depth: 12 00:00:00 215kN 28...g6 29.f5 Ng4 30.Ng1 Qh4 31.Rg2 = (0.24) Depth: 12 00:00:00 368kN 28...b6 29.Ng1 Qh4 30.Rg2 Rcb8 31.f5 Re8 = (0.08) Depth: 12 00:00:01 465kN 28...Ng4 29.Rg2 f5 30.Ng1 Qh4 31.Nf3 = (-0.03) Depth: 12 00:00:01 683kN 28...Ng4 29.Rg2 Ne3 30.Bxe3 dxe3 31.Rg3 Qe6 32.Rfg1 g6 33.Qe2 a4 34.f5 Qd6 35.Qxe3 ² (0.44) Depth: 15 00:00:06 3689kN 28...Kh7 29.Rg2 Nh5 30.f5 Rf8 31.Ng1 Qh4 32.e5 ² (0.26) Depth: 15 00:00:11 6794kN 28...Kh7 29.Rg2 Nh5 30.Ng1 Qh4 31.f5 Bd6 32.Nf3 Qh3 33.e5 ² (0.26) Depth: 17 00:00:37 22145kN 28...Kh7 29.Rg2 Nh5 30.Ng1 Qh4 31.Rg4 Qf6 32.Nf3 Re8 33.Rfg1 Qd6 ² (0.32) Depth: 18 00:01:38 55004kN 28...Kh7 29.Rg2 Nh5 30.Ng1 Qh4 31.e5 Bf8 32.Qf3 b6 33.Rg4 Qe7 34.Qe4+ Kh8 35.f5 ² (0.42) Depth: 19 00:03:58 130129kN 28...Re8 29.Rg2 Kh8 30.Rg3 Qc8 31.Ne5 Kg8 32.f5 ² (0.38) Depth: 19 00:07:20 240382kN 28...Re8 29.Rg2 Bf8 30.f5 Nd7 31.Rg3 ² (0.49) Depth: 20 00:09:32 311679kN 28...Re8 29.Rg2 Bf8 30.Rg3 Qd7 31.f5 Kh7 32.Rfg1 b6 33.R3g2 ² (0.49) Depth: 21 00:20:47 672721kN 28...Nd7 29.Rg2 Kh8 30.Rfg1 Rg8 31.Rg3 Qe6 32.Qd2 g5 33.fxg5 hxg5 34.Nxg5 Bxg5 35.Rxg5 Qh3 ² (0.45) Depth: 21 00:29:42 960629kN (Smith, Stevenson Ranch 19.04.2002) 28...Rf8, after which Smirin made the most of the stranded queen, shows up for a brief moment on my system at depth 6, and never again. Either Junior blitzed this instantaneously at Smirin, or it's something out of the hash. It churned through a lot of moves, but it looks like 28...Kh7 or 28...Re8 would have been played at a longer time control. Neither one provides the queen an escape route after White plays f5, as f6 remains occupied. Finally, at nearly 30 minutes (5-6 on their machine?), it finds the exact line recommended in the notes at Chessbase through move 30. Perhaps the actual author of this note is Junior itself, picked up by the annotator when Junior was analyzing during the postmortem.
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.