<1.e4 b6 2.d4 Bb7 3.Nc3 Nf6 4.Bd3 e6 5.Nf3 Bb4 6.Qe2 d5 7.exd5 Bxd5 8.0-0 Bb7 9.Bg5 Be7 10.Bb5+>

Not seeing a good way to refute Black's opening, I decided to try a dubious one. The thought is that Black doesn't want to open the position with ...c5 until he is better developed. By encouraging him to play ...c6, his Bb7 will be blocked until he plays ...c5 whether he likes it or not.

<10...c6 11.Bd3 0-0 12.Rad1 h6 13.Bxf6 Bxf6 14.Qe4 g6 15.Ne5 Qe8 16.Rfe1 Na6>

This is always a difficult decision for me--cash in for a pawn or continue the build-up? As too often happens, I made the wrong decision, especially since White can't hold the pawn in the way I play it.

<17.Bxa6 Bxa6 18.Qxc6 Bxe5 19.Qxe8 Bxh2+ 20.Kxh2 Rfxe8 21.d5 exd5 22.Nxd5 Rxe1 23.Rxe1 Rd8 24.Nc7>

click for larger view

Here I offered a draw, figuring the momentum was shifting and that Black had the better endgame anyway. Black decided to play on, rightfully in my opinion.

<24...Bb7 25.Re8+>

A bishop vs. a knight generally has the advantege in an open position with pawns on both sides of the board, and I figured that having a long-range piece like a rook on the board would only make it worse.

<25...Rxe8 26.Nxe8 f5>

Perhaps 26...Kf8, activating the king before committing the pawns, is a better idea. As the game goes, Black tries to get his kingside majority going, but the pawns turn out to be vulnerable when White uses the time to post his king aggressively.

<27.Kg3 Kf8>

I was much more afraid of 27...g5, when I couldn't see an effective way to get the king into play quickly. Another thought is 27...f4+, so that 28.Kxf4 Bxg2 gives Black a passed pawn.

<28.Nd6 Bd5>

Looks obvious, but maybe 28...Bc6 is better. The text gives White a chance to push his majority with tempo, and psychologically prepares the next dubious move. Also, 28...f4+ was still possible.

<29.c4 Be6>

I think the bishop should stay on the long diagonal, but it's hard to admit that you've lost a tempo.

<30.Kf4 Ke7 31.Ke5 a5 32.b3 f4>

Trappy, but it gives White's knight what turns out to be a useful square on e4. In addition, Black soon has to play ...g5, when the pawns further weaken themselves by advancing.

<33.f3>

It felt odd to put my pawns on the same color as Black's bishop, but it could only attack them from the rear and can't get there from here.

click for larger view

And now Black offered a draw, which I felt justified in refusing.

<35.Nf6 Be6>

White was threatening two pawn-winning forks, but this gives in a bit too easily. After 35...Kd8 White would have some chances by trading on d7 immediately; both sides promote, but White winds up a pawn or two after all is said and done. However, if he first finds 36.Ng8 h5 37.Nf6 h4, Black's pawns are weakened further and the math turns in White's favor: he will need only eight moves to promote, while Black will need ten. (Play it out for yourself; this post is too long already.)

<36.Nd5+ Bxd5 37.cxd5>

37.Kxd5 might win after some arcane zugzwang manuevers, but this seemed easier to calculate. Black's king has to keep an eye on the d-pawn and can't defend the kingside, and he will always a move or two late on the queenside.

<37...b5 38.Kf5 Kd6 39.Kg6 Kxd5 40.Kxh6 a4 41.bxa4 1-0>

click for larger view

OK, I'll give one line to show how close this might have been: 41...bxa4 42.Kxg5 Ke5 43.Kg4 a3 44.Kg5 Kd4 45.Kxf4 Kc3 46.g4 Kb2 47.g5 Kxa2 46.g6

click for larger view

and now all Black has to do is move his king and promote the pawn--except that the king has no good squares! 46...Kb3 allows White to queen with check, and 46...Kb2 or 46...Kb1 allow checks on the diagonal or rank respectively, followed by trading queens and coronating the last pawn.

Very close! Of course, White could have made it a lot easier on himself with 43.a3 when Black is a move behind the above variation, but what fun would that be?

fictionist (1689) - phonybenoni (2034)
Challenge [Queen Alice] 23.05.2008

<1.e4 e5 2.Nf3 Nc6 3.Bb5 f5>

My second line of defense against the Spanish, in which I indulge when too many people are playing the Exchange Variation or my opponent casully mentions his extraordinary record against the Marshall.

<4.Nc3 Nf6 5.Qe2 Nd4 6.Nxd4 exd4 7.exf5+ Be7 8.Ne4 0-0 9.0-0 c6>

For better or for worse, this seems to be my own idea; Opening Explorer mentions only 9...Nxe4. I didn't see why I couldn't play it, and the position begins to look more like a Bird's Defense.

<10.Bd3>

The bishop looks awkward here, but it might be better than an eventual retreat to b3 once Black gets ...d5 in.

<10...Nxe4 11.Qxe4>

11.Bxe4 d5 retrieves the gambit pawn at once. Of course, whether White should be trying to hold the pawn on f5 is another question.

<11...Bf6 12.c4>

This surprised me, and I think it's a good idea. Black can hardly free white's game by taking en passant, but now Black won't be able to play ...d5 as effectively.

<12...d6 13.b4 Be5 14.Bb2 Bxf5 15.Qe2>

click for larger view

I thought Black was fine here, with kingside pressure compensating for the silly queenside pawns. Unfortunately, I conceived a brilliant combination to prove it which my opponent anticipated and was ready for.

<15...Bxh2+ 16.Kxh2 Qh4+ 17.Kg1 Rae8 18.Qxe8>

I was anticipating 18.Qf3 Be4 "regaining the piece", but even then White can give up the queen with 19.Bxe4 Rxf3 20.Bxf3 and have the same material imbalance as in the game.

<18...Rxe8 19.Bxf5>

I was now sure that I was lost, with queen against rook and two bishops. In fact, Black still has chances to hold the balance by grabbing queenside pawns, and I follow the right path for a while.

<19...Qf4 20.Bh3 Qxd2 21.Rab1 Qxb4 22.Bxd4 Qxc4 23.Bxa7 b5 24.Rbe1 Rxe1 25.Rxe1 d5>

An interesting idea is 25...Kf7 26.Be6+ Qxe6 27.Rxe6 Kxe6, when the queenside pawns might give White's remaining bishop some headaches. I eventually saw this idea when it was too late.

<26.Bb8>

click for larger view

<26...Qxa2>

Suicide. After 26..,Kf7 or some other reasonable move keeping the queen in the game, it's still a hard fight.

<27.Be6+ Kh8>

Here I was hoping that White would get fancy with 28.Bxd5 Qxd5 29.Re8+ Qg8 30.Rxg8+ Kxg8, when we have a form of the ending metioned a bit earlier. But his simple reply snuffed out all hope.

<28.Bf7 1-0>

click for larger view

Analysis by Rybka 2.3 mp 32-bit (21-ply):

1. = (0.21): 18.Qf3 Be4 19.Bxe4 Rxf3 20.Bxf3 Rf8 21.Be2 Rf6 22.f4 Rh6 23.Bf3 Qh2+

2. = (0.00): 18.Qxe8 Rxe8 19.Bxf5 Qf4 20.Bh3 Qxd2 21.Rab1 d3 22.Rfd1 Qc2 23.Bf5 Rf8 24.Be6+ Kh8

3. (-1.49): 18.Be4 Bxe4

4. (-1.49): 18.g3 Rxe2

5. (-1.49): 18.Rae1 Rxe2

And just before the critical <26...Qxa2>. I'll hand it to <Rich> though. He found the best moves at the end.

click for larger view

Analysis by Rybka 2.3 mp 32-bit (21-ply):

1. (0.42): 26...Kf7 27.Bd6 d4 28.Be6+ Qxe6 29.Rxe6

2. ± (0.72): 26...Qc5 27.g3 Kf7 28.Be6+ Kf8 29.Bf4 Qe7 30.Re5 Qc5 31.Bd7 Qc2 32.Re1 d4 33.Be6

3. ± (0.72): 26...Qb4 27.Rc1 g5 28.Be6+ Kg7 29.Rxc6 Qe1+ 30.Kh2 d4 31.Bb3 Qe8 32.Rc7+ Kh6 33.Rb7

4. ± (0.76): 26...g5 27.Be6+ Kg7 28.Be5+ Kg6 29.g4 Qb4 30.Re2 Qc5 31.Bf5+

5. ± (1.01): 26...g6 27.Re7 Qb4 28.Rd7 Qd4 29.Kh2 g5 30.Bf5 Qf6 31.Bxh7+ Kf8 32.Ba7 Ke8 33.Rb7

click for larger view

Analysis by Rybka 2.3 mp 32-bit (20-ply):

1. (-0.44): 15...Qg5 16.Bxf5 Qxf5 17.d3 Rae8 18.Qd1 Re6 19.Bc1 Qg6 20.Rb1 Ref6 21.Qc2 Rf3 22.Rb3

2. = (-0.20): 15...a5 16.Bxf5 Rxf5 17.a3 Qf8 18.d3 Re8 19.Qg4 Rf4 20.Qh5 Re6 21.g3 Rf3 22.Rad1

3. = (-0.15): 15...Qf6 16.f4 Bxf4 17.Rxf4 Bxd3 18.Rxf6 Bxe2 19.Rxd6 Bxc4 20.Rxd4 Bd5 21.Re1 Rae8 22.Rxe8

4. = (-0.11): 15...Qh4 16.g3 Qg5 17.f4 Bxd3 18.Qxd3 Rxf4 19.Rxf4 Bxf4 20.Bxd4 Bxd2 21.a3 Re8 22.Rd1

Actually, I have to say that I'm often very wary of computer evaluations. If one line is a +0.21 and the second line a +0.42, I think it's ridiculous to insist that the second line is "better". Not only is the difference insignificant mathematically, but there might be psychological and other human reasons to prefer the first line. Let me give you a very bad example:

phonybenoni (2024) - kelton (1675) [B01]
AUTO-OPEN-1043 [Queen Alice] (2), 30.05.2008

<1.e4 d5 2.exd5 Qxd5 3.Nc3 Qd6 4.g3 Nf6 5.Bg2 g6 6.d4 Bg7 7.Bf4 Qb6 8.Nge2 Qxb2 9.Rb1 Qa3 10.Bxc7 Nbd7 11.Qd3>

click for larger view

A sleazy trap into which Black falls hook, line and sinker.

<11...0-0 12.Rb3 Qa6 13.Qxa6 bxa6 14.Bxa8 Nb6 15.Bxb6 axb6 16.Rxb6 Bf5 17.Rxa6 Bxc2 18.Kd2 Bf5 19.Nd5 e6 20.Ne7+ Kh8 21.Nxf5 gxf5 22.Bc6 Bh6+ 23.Kd3 Ng4 24.Rf1 Nxh2 25.Rh1 Ng4>

click for larger view

<26.Rxh6!>

I'm sure Rybka would not consider this the best move by a long shot, but it absooutely was for me. White now has a completely clear plan: trade rooks and march the a-pawn. Black's king is too far away to do anything, and knights always have trouble with rook pawns even if they're not on the other edge of the board. There was no need to calculate anything, which means a lot fewer blunders on my part.

<26...Nxh6 27.Ra8 Rxa8 28.Bxa8 Kg7 29.a4 Ng8 30.a5 Ne7 31.Bb7>

Had Black played 30...Nf6, then 31.Bc6 would have been a similar example "cut him off at the pass". That's why I didn't move the bishop earlier (I would have loved to play Bh1 at some point).

<31..Kf6 32.a6 e5 33.dxe5+ Kxe5 34.a7 Ng6 35.a8Q h5 36.Qa1+ Ke6 37.Nf4+ Ke7 38.Nxg6+ fxg6 39.Qg7+ Kd6 40.Kd4 g5 41.Bc8 h4 42.Qd7# 1-0>