I was unfamiliar with this move, but it gets 20 pages in Nunn's The Complete Najdorf: 6.Bg5. 8.Qf3 Qc7 9.0-0-0 b5 10.e5 As it turns out, this is the sharpest method at White's disposal. It's not the most popular, as the fireworks end up petering out into a draw in the major branches - or so theory thinks. At any rate, the game is about to enter some incredible complications. 10...Bb7 11.Qh3 dxe5 12.Nxe6 fxe6 13.Qxe6+ Be7  At any rate, White has a tough choice here. My initial idea was 14.Nxb5 axb5 15.Bxb5 and Black resigns. Then, I saw a strange idea for Black: 15...0-0-0. At first it looks like just a desperate way to keep the game going, since after 16.Qxe7 White must just have a winning position, but what concerned me was that Black could play 16...exf4 and still be up material. I looked at 17.Qxg7 but then 17...Rhg8 and White's pieces seem pretty awkward. I hate to reject a move like 14.Nxb5 because of a move like 15...0-0-0, but in the end I concluded that Black stood better, and since after 14.Bxb5 White has at least a perp and several promising-looking ways to play for more, I played 14.Bxb5. As it turns out, 14.Bxb5 is in fact better than 14.Nxb5, as it, along with 14.Bxf6, are both thought to lead to draws by perpetual, while 14.Nxb5 is in fact thought to favor Black - but watch how! First of all, after 14.Nxb5 axb5 15.axb5, 15...0-0-0 is in fact thought to be inadequate for Black (as really it should be by the looks of it), since after 16.Qxe7 exf4 17.Qxg7 Rhg8 White plays 18.Bxd7+ Nxd7 19.Bxd8, hitting the queen, with a much better position. This was actually pretty simple, and I think I should have seen it. Far more shocking is the reason why 14.Nxb5 is bad: 14...axb5 15.Bxb5 Be4!! It seems ridiculous to allow 16.Bxd7+, but then after 16...Kf8 all of a sudden White's king is flooded with threats: 17.c3 Ba3 18.Qxe5 Qb6 19.Rd2 Ra5, or 17.Rd2 Bb4 etc. In fact, 16.Bxd7+ isn't even White's best move. According to theory, he should opt instead for 16.Rd2, but then after 16...Kf8 17.Bc4 Bg6 he is not thought to have sufficient compensation for his two pieces. So, right move, wrong analysis - though practically speaking 14.Nxb5 would have won since there is no way that my opponent would have found 15...Be4. 14.Bxb5 axb5 15.Nxb5 Qc6 16.Nd6+ Kd8 17.fxe5  17...Nd5 As it turns out, Black had another surprising possibility here: 17...Kc7, which seems to lead to an immediate forced draw after 18.Qxe7 Rxa2 when White apparently can't avoid the checks starting with ...Ra1+ and ...Qd5+ etc. The winning try 18.Kb1 doesn't seem to leave White with enough compensation after 18...Bf8. According to Nunn 17...Nd5 also leads to a forced draw, but in this line White does have several ways to try to play for a win which seem somewhat promising to me, so 17...Kc7 is probably at least more accurate. 18.Bxe7+ No alternatives are given by Nunn here. I looked briefly at 18.Nxb7+ Qxb7 19.Rxd5 Bxg5+ 20.Kb1 but then Black has 20...Re8 - probably among other things. 18...Nxe7 19.Nf7+ Ke8  At the board, my choice was between 20.Qxc6, 20.Nd6+ Kd8 21.Nxb7+ Qxb7 22.Rxd7+ (I don't see why Nunn completely ignores this line - I've checked it with my computer), and 20.Rd6. The first two lines (20.Qxc6 and 20.Nd6+ Kd8 21.Nxb7+) aren't very complicated tactically but lead to extremely unbalanced positions in which judgment is required. I decided that while after 20.Nd6+ Kd8 21.Nxb7+ Qxb7 22.Rxd7+ Qxd7 23.Rd1 White shouldn't be at a serious risk of losing, Black also wouldn't be in much danger of losing. I've found these sorts of crazy imbalances to often lead to drawish positions - perhaps Black sets up an impregnable fortress, or else he really starts using his three pieces to terrorize White's pawns but White finds a perpetual, etc. For what it's worth, Fritz likes White slightly here - but that doesn't mean much. I haven't changed my mind about this - I would still expect the result of the game to be a draw here - but who knows? 20.Qxc6 is a little more double-edged. Black has several ways to win two pawns back and keep the knight stuck on h8, and while rook and two pawns vs two minor pieces tends to be good for the rook with queens off the board, it seems to me that here White's stranded knight on h8 should more than make up for that. My favorite line for Black is 20...Nxc6 21.e6 (21.Nxh8 Ncxe5 and White's knight has problems) 21...Nc5 22.Nxh8 Ke7 23.Nf7 and maybe now 23...Rxa2 - but other lines are plausible as well. Nunn gives one game here where Black did not go after the pawns - Hector—Nedobora, which went 20...Bxc6 21.Nxh8 Bd5 22.e6 Bxe6 23.Rhe1 Ra6 which Nunn gives as unclear. I'd probably rather have White here - and of course Fritz would too. Still, it seems hard to believe that White has serious winning chances in this line, especially given all of Black's options. This leaves 20.Rd6, which is very complicated tactically. It it turns out, it should lead to a position which, while not straightforward to assess, is at least somewhat better for Black. 20.Rd6 Nc5 21.Qc4 Forced, since 21.Qg4 loses to 21...Qxg2, ie.22.Qd1 Kxf7.  At the board I thought that this was forced and that I was much better, but in fact Black has two serious alternatives, one unclear but probably good for White and the other also unclear but probably good for Black. 21...Ba6 also seems to be better for Black. These two major alternatives are 21...Qe4 and 21...Qa4, the point being that in both cases 22.Qxc5 fails to 22...Qf4+ 23.Kb1 Qxf7. 21...Qe4 does in the end seem to turn out inferior after 22.Rd4 Qe3+ 23.Kb1, the main line according to my analysis on the computer being 23...Nf5 24.Nxh8 Nxd4 25.Qf7+ Kd8 26.Qf8+ Kd7 27.Qxc5 with an edge for White. 21...Qa4, however, forces 22.Qxa4 Rxa4 23.Nxh8 when Black seems to at least have improved on the 20.Qxc6 Bxc6 21.Nxh8 lines. I am not sure how to actually assess this position, but offhand I'd rather be Black. Certainly, if I'd seen that 21...Qa4 forces 22.Qxa4, that would have been the end of 20.Rd6.21...Ba6, however, while more complicated, seems even better. 22.Qg4? 22.Rxc6 Bxc4 23.Nd6+ Kd7 24.Rxc5 Rxa2 was forced, and while this position too is not easy to assess, I am almost certain that Black at least has all of the winning chances. Another interesting but ultimately losing try is 22.Qd4. 22...Nf5 23.Rxc6 Nxd4 24.Nd7+ Kd7 25.Rxc5 would then be a definite improvement for White over the 22.Rxc6 line since White's a-pawn is shielded, but the problem is that Black just throws in 22...Nd3+ 23.Kb1 Nf5 and wins since after 24.Rxc6 Nxd4 25.Rxa6 Rxa6 26.Nxh8 Nxe5 he wins the stranded knight while 24.Qg4 runs into 24...Bc4 25.a3 Qb5. After 22.Qg4, the following position has arisen:  Prior to playing 20.Rd6, I spent a ton of time - probably at least 20 minutes - "refuting" 22...Nd3+ here. I believed that I had done so, but in fact 22...Nd3+ 23.Kb1 Nf2 wins for Black. I had planned 24.Qf4, but this actually meets two refutations: 24...Qc4 25.Qxf2 Qxf7, as well as 24...Qxg2, since after 25.Re1 Nh3 White can't continue to defend f7. 24.Rxc6 also loses, since after 24...Nxg4 25.Rxa6 Rxa6 26.Nxh8 Nxe5 White will again lose his knight. Best seems to be 24.Qxg7 Rg8 25.Qxg8+ Nxg8 26.Rxc6 Nxh1 27.Rc7, with another crazy material imbalance, this time White having five pawns for two pieces, but I imagine that White is just lost here. There was one line that I calculated at the board that has withstood the computer's assault: that 22...Nd3+ 23.Kb1 Nxe5 loses to 24.Qxg7 Nxf7 25.Rxc6 Nxc6 26.Re1+ Ne7 27.Qf6 Ra7 28.Qb6. Even better yet, it takes a while to convince the computer that White is winning here. 23.Qxg7?? This final oversight ends the game. White should play 23.Qxe4 Nxe4 24.Nxh8 Nxd6 25.exd6 Nf5 26.d7+, escaping into a piece vs 3 pawns ending which I imagine he should be able to draw with accurate play - though who really knows? 23...Rg8 24.Qf6 Rxg2 My original idea here was 25.Rd2 Rxd2 26.Nd6+ Kd7 27.Nxe4. The problem is 26...Rxd6. This is the sort of line that needs to be double-checked three times - I don't care what Kotov says. 25.Qh8+ Rg8 26.Qf6 Qxh1+ 27.Rd1 Qxd1+ 0-1. The crazy thing is, I thought I saw a ton in this game and that my opponent was just really lucky. It's kind of scary running these sorts of games through a computer. I guess this sort of stuff is just too complicated for 40/2. |