Author: Vincent Lejeune
Date: 13:52:38 04/24/02
[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 0.01 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.