Author: Chris Carson
Date: 14:53:36 04/24/02
Go up one level in this thread
On April 24, 2002 at 16:52:38, Vincent Lejeune wrote:
>[Event "Internet challenge"]
>[Site "?"]
>[Date "2002.04.24"]
>[Round "?"]
>[White "Ilya Smirin"]
>[Black "Hiarcs 8"]
>[Result "1/2-1/2"]
>[ECO "A30"]
>[WhiteElo "2702"]
>[Annotator "Frederic"]
>[PlyCount "119"]
>
>1. c4 c5 2. Nf3 Nf6 3. g3 d5 4. cxd5 Nxd5 5. Bg2 Nc6 6. d4 cxd4 7. Nxd4 Ndb4 8.
>Nxc6 Qxd1+ 9. Kxd1 Nxc6 {Evan: Eureka, open game!! Mig: Not a bad idea against
>a comp. Get the queens off the board and you almost never get mated!} 10. Nc3
>Bd7 {Mig: Symmetrical pawn structure, equal material... Are they just going to
>agree to the draw now? :-)} 11. Be3 e5 {Mig: White has chances to do something
>with his minimally better development, but I'm not sure he will. A weak pawn
>on c6 is one thing, but giving the computer the advantage of the bishop pair
>would not be good. The bishop pair in an open position is worth as much as an
>isolated pawn or two. Smirin has to count on HIARCS making mistakes in this
>simple position.} 12. Rc1 Rd8 {Mig: Since computers can't plan as such, they
>often wander into trouble when they can't find a clearly superior line of
>analysis. But this position may be too simple for that to happen. Time: 54:1.
>01} 13. Ke1 {Ardee: The most boring game of the series by far. Mig: Other ways
>computers can lose these positions include getting overextended pawns. For
>example, f5-e4 or something, then the pawns could be lost in an endgame. And
>we all know how HUMANs can lose these positions to computers. Fine, good, okay,
>fine, OOPS, MY ROOK! White's bishops create a lot of pressure on the black
>queenside. The black rook might have been better left ona8 in the long run.}
>13... Be7 14. f4 {Mig: f3 would just lock up the g2 bishop, but the white king
>needs to make way for the h1 rook. So f4, which also challenges black in the
>center. Time: 53:55} 14... exf4 15. gxf4 {Twinzen: Why not Bxf4? Mig: It's not
>so much why not Bxf4, but why. It wouldn't do much at all, and would leave an
>isolated e2 pawn.} 15... O-O 16. Nd5 {Mecking: Mecking says Bd6 also.} 16...
>Bd6 17. Kf2 {Mig: Note that gxf4 created a potential passed e-pawn. Far, far
>away, but it's an imbalance that a human understands, while the comp can't
>imagine that e2 pawn becoming a queen on move 48.} 17... Be6 18. Rhd1 {
>Mig: White would like to 'encourage' black to give up the bishop pair. Then
>the black a and b pawns will be hard to defend in the long run.} 18... Rd7 {
>Time: 45:50. Ardee: Nc3 followed by Rg1 for White? Mig: HIARCS tries to set up
>some tactical tricks on the d-file. I like the plan. White must either double
>on the d-file himself or leave it to black. b4 looks risky for ahuman!} 19. Nc3
>{Mig: Protecting d1 with the knight, but losingtime. Any advantage white had
>is gone now.} 19... Rfd8 {
>GM Mecking: I agree with you Mit. Black is ok in this position.} 20. a3 Ne7 {
>Mig: White still has queenside pressure with his bishops and isn't panicking
>here. But black is headed for f5. Bxa7 Bxf4 doesn't look like much fun for
>white, but at least the pressure on the d-file would be broken and Smirin
>could probably draw that ending. 21.Rd3 would also block the black pressure on
>the d-file somewhat, but doesn't do much about ...Ng6, hittingthe f-pawn.
>Playing the king up is another try. Mecking: I have to go. I enjoyed my first
>visit to the Fritz server. Time: 41:47} (20... Ne7 21. Rxd6 {Mig: That's an
>interesting exchange sac, but I think it's a bit early to panic. The two
>bishops would provide some compensation, but white has a few 'normal' tries
>first.}) 21. Be4 Nf5 22. Bxf5 Bxf5 {
>Mig: It was either give up the bishops or give up a pawn. Time: 32:45} 23. Rd4
>a6 {Mig: Now White has the goal of winning a bishop for his knight, even if it
>costs a pawn. If he can get rid of the black dark-squared bishop then he will
>have very good drawing chances thanks to the opposite colored bishops. We'll
>see if HIARCS knows to avoid this. In general the computer will try to keep
>the bishops anyway. The rooks are destined to be swapped thanks to their
>placement on the d-file. He played Rd4 to play Ne4, and a pair of rooks will
>go after Be7 Rxd7.} 24. Na4 {Mig: Hmm, what about Bxa3?} 24... Be7 {
>Mig: avoids back-rank mate tricks.} (24... Bxa3 25. Rxd7 Bxd7 26. bxa3 Bxa4 {
>is quite playable. Hard to win, however.}) 25. Rxd7 Bxd7 26. Nc3 {Mig: HIARCS
>didn't take the bait! I asked earlierif it would understand the opposite
>colored bishops drawing potential. It just might!} 26... Bh4+ {Mig: Always
>attacking chances with the bishops on theopen board. And excellent endgame
>winning chances because pawns on both sides of the board.} 27. Kg1 {Time: 25:42
>} 27... f5 {Mig: Black has two pawn islands versus white's three, Black has
>the two bishops in an open positions with pawns on both sides of the board.
>That's two small advantages that are pretty big when combined.} 28. Rd1 Bc6 29.
>Rxd8+ Bxd8 {Mig: The remaining question is if HIARCS has the technique to win
>this without a Smirin blunder. Human praxis says black should keep the rooks
>on the board and make sure the white knight has no advanced posts.} 30. Bd4 Kf7
>31. e3 {I thought black had a serious advantage a few moves ago, but it's
>going to be harder to win this in my opinion. TheVish: h6 and g5, Black should
>be much better. Time: 23:40} 31... h6 32. Ne2 {Mig: Black has to find an entry
>point for his king and then give up one of the bishops at the key moment to
>penetrate. It will be interesting to see if HIARCS can do something so
>far-sighted. So far black is 'planning' well. Kingside pawn-push coming.} 32...
>g5 {TheVish: fxg5 is forced, otherwise g4 must be seriously good for black.}
>33. Kf2 Bc7 {TheVish: Why not g4? Why would a computer not play g4, Fred?} 34.
>Bc3 {Time: 19:39} 34... Be4 35. Kg3 {Mig: Another chance to play ..g4. I don't
>see how black can hope to win without playing that eventually.} 35... g4 36.
>Bd4 {Mig: Smirin is saying, "Nyah nyah, stupid computer can't figure out a
>winning plan if I play prophylaxis."} 36... Kg6 {Time: 19:35} 37. Kf2 {
>TheVish: ...h5} 37... h5 {Sith: Oh the Humanity!!} 38. Bc3 {
>TheVish: Smirin wants to just wait eternally after h4} 38... Bd8 {Mig: So here
>we arrive at one of the traditionallly weakest areas in computer play. Endgame
>planning. If the computer can't find a solid eval gain down one line it will
>just maintain the status quo.} 39. Bd4 {TheVish: Bd4 and Nc3 is a fortress.
>Maybe black can play h4-h3, force the knight to g3 and then take the king to
>d3} 39... a5 40. Bc3 a4 {Shay says Simirn is gazing at the monitor. Mig: At
>least black can never lose with the white king stuck defending the kingside.
>Otherwise I'd be worried black might lose with all these pawn pushes..} 41. Nd4
>{TheVish: Bh4+!} 41... Bh4+ 42. Kg1 {Time: 16:30} 42... Be7 {
>TheVish: This could go on for ever!} 43. Kf2 Bd5 44. Kg3 h4+ 45. Kf2 Be4 46.
>Kg1 g3 {Time: 14:19. TheVish: It's still a long way to resign.} 47. h3 {
>TheVish: Black has to play Bc5, then White can play Bb4, followed by Ne2-c3
>and it's a fortress. Bc5 Bb4 is adraw} 47... Bc5 {time: 11:28} 48. Bb4 {
>Mig: Smirin has avoided the blockade before. He seems to be daring the
>computer to find a win. Of course if the pieces come off the pawn endgame is
>completely lost for white.} 48... Bxb4 49. axb4 {
>TheVish: if Bd3 then Nf3, otherwise Ne2-c3 and Black's king can never enter}
>49... Kf7 {Mig: This is why I always say comp evals in endgames are worthless.
>They should be binary. WIN or DRAW. None of this -2.39 junk in an endgame.
>TheVish: It's over!} 50. Nb5 {TheVish: well, ok all roads lead to c3} 50... Ke6
>51. Nc3 {TheVish: phew! Mig: That's what always drove me nuts commentating
>when people in the audience are saying, "but ChessSpank 9001 says it's +2.31!"
>in an endgame that was a book draw.} 51... Bc2 52. Kg2 {Time: 11:25} 52... Kd6
>53. Kg1 Kc6 54. Kg2 {
>Mig: Hmm, a variation of the dreaded Kh1-Kg1 attack: The Kg1-Kg2 Attack!} 54...
>b5 55. Kg1 Bd3 56. Kg2 {Time: 11:22} 56... Be4+ 57. Kg1 Bc2 58. Kg2 Bd3 59. Kg1
>{Mig: Hmm, this position looks like lunch to me.} 59... Be4 60. Kf1 {
>DRAW AGREED} 1/2-1/2
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.