Between 2010 and 2011 I played a two-game match against Rosario DiPeri, one of SchemingMind's Senior Masters. He beat me easily with White and we drew the other game, which was a fascinating struggle in the Sveshnikov Variation of the Sicilian defence.
At one point he chose to go into what I considered a dubious line, but I couldn't exploit it and he defended formidably. After the game I decided to investigate it further and I analysed the game with a strong computer, and came to the conclusion that this continuation loses by force. Here are the findings.
Kajetan Wandowicz -- Rosario DiPeri
1.e4 c5 2.Nf3 Nc6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 e5
6.Ndb5 d6 7.Bg5 a6 8.Na3 b5 9.Nd5 Be7 10.Bxf6 Bxf6 11.c3
- This is the heavily analysed Sveshnikov variation, and for many moves we will be following theory.
11...O-O 12.Nc2 Bg5 13.a4 bxa4 14.Rxa4 a5 15.Bc4 Rb8 16.Ra2 Kh8 17.Nce3 g6?! +/-
- One of the main lines.
- And this is the continuation that is the subject of my article. The purpose of this move (and of course of the previous one, 16... Kh8) is to achieve the break ...f5. However, this leads almost by force to a position that seems to be objectively lost for Black. Winning it is another story though, and this line is certainly playable in over-the-board chess, as evidenced by Grandmaster practice. The position quickly becomes a total chaos in which anything can happen.
18.h4 Bxh4?! +-
- (17...Bxe3! is in my opinion the right way to play for ...f5, and has been chosen by GM Veselin Topalov (in 2006, when he was at the top of his form with a rating of 2801!) among others. In correspondence chess, this move has attracted many very strong players, and has scored significantly better than 17... g6. 18.Nxe3 Ne7 19.O-O (or 19.b3 f5) 19...f5 and Black has equalised, though the game is far from finished, as the position is extremely complex.)
- This move was played by Grandmasters Vladimir Kramnik (a great Sveshnikov expert earlier in his career) and Loek van Wely, among others, but I think it loses for Black. That's one of the differences between over-the-board and correspondence chess! The game is so complicated, that even the World's top players are bound to make some mistakes here, and even the slightest one can be deadly for either side. With perfect play, however, I think that White wins. In this game it took a serious mistake on White's part for Black to draw, even though Black was way stronger. On the other hand, once Black has committed to 17... g6, it's difficult to suggest an alternative.
19.g3 Bg5 20.f4 exf4 21.gxf4 Bh4+ 22.Kf1 f5 23.b4 fxe4 24.Rah2 g5 25.Ke2!
- (18...Bxe3 may be considered; this is the last point at which anything can be done)
- Immediately sidestepping the pin on the f-file, thus depriving the black knight of the square e5.
- (25.b5?! +/- is most commonly chosen, but the text is stronger. That is not to say that 25.b5 leads to an advantage to Black, far from it, but maybe he has more chances to hold the game in the resulting position -- thus, the move should be regarded as dubious in correspondence chess. 25...Ne5 26.Qd4 Rb7 27.Rxh4 gxh4 28.Ke2 Re8 29.fxe5 Rxe5 30.Nf4 Qg5 31.Qxd6 Bg4+ 32.Ke1 Qxf4 33.Qd8+ Kg7 34.Qg8+ Kh6 35.Rxh4+ and Black resigned in the rapid game Kramnik - van Wely, Monte Carlo 2005. This was of course only a rapid game, but it shows that the position is complex enough for even a very strong Grandmaster to go wrong.)
- This is the novelty (we're at move 25 -- that's how heavy the Sicilian theory is!), but it is absolutely forced, as White would now play 26.b5, and the black knight on c6 would become quite sad. In the previous two games Black had missed this threat and played something else, to which in both cases White replied 26.b5 and subsequently won the game.
- Threatening Rxh4, which Black can't stop.
- (It was of course impossible to move the bishop: 26...Bxg5?? 27.Rxh7+ Kg8 28.Ne7# is an elegant checkmate.)
- White now has a piece for three pawns, and is threatening the same checkmate as in the note to the previous move.
- Parrying the threat.
- This is tempting, because it recaptures the pawn, threatens the knight at c6 and opens the great bishop at c4; surely a move that accomplishes all this can't be bad?! Unfortunately, it is, throwing away practically all the advantage. White should renew his threat of b5. Now Black has to play precisely, but he can just hold the game.
- (28.cxb4! and b5 is threatened again. Note that Black can't capture on b5, as White would recapture with his own knight and the rook would be forced to abandon b7, where it is needed to defend against the checkmate threat. 28...Qg3 Threatening Qf2+ or Rf2+, but it's not enough. Only computer analysis reveals it though! That's why this variation is playable over-the-board, or even in correspondence chess without computers -- who would allow this with white? 29.Qa1+! Ne5 30.Kd1!! The only move, as revealed by detailed computer analysis. But who would find it without a computer, allowing the opponent's pieces so close to the open king? If Black doesn't pose any threats now, white will simply play Qd4, centralising his queen, and take the e4 pawn with an easy win. White king is amazingly safe here, and can go to the c-file if it is bothered. 30...Be6 defending the diagonal 31.Qd4 Ra8 32.Qxe4 White now is a piece up and all he has to do is trade down to release the tension. Black's rooks look so scary though on the abandoned queenside, that it is difficult to believe White's king is completely safe. First though, White threatens Rxh7, so Black has to defend. 32...Bg8 33.R4h3 Qf2 Now it is Black who threatens checkmate! 34.Nc2 Covering a1. 34...Rba7 Renewing the threat of Ra1+ mating. 35.Be2! +- And White is completely safe. Now it is enough to slowly unravel and win with the extra material. Finding the required moves however, and letting my king walk on a tightrope in the process, proved too difficult this time.)
- The only move. Black now gets the initiative, using the fact that White abandoned the defence of his e3 knight to threaten Qxe3 checkmate. From now on, both sides play logically and without major imprecisions, and the game ends in a draw.
29...Rxe3+ 30.Nxe3 Rb2+ 31.Nc2
- Now that White captured the b-pawn, Black's rook has an open file, so White has to defend from a deadly check threatened on b2.
- (After the game, Rosario told me that he expected 31.Qd2 Rxd2+ 32.Kxd2 which he regarded as a little better for White. In my opinion, even if White is better, it's only very slightly, and I was a little afraid of Black's two pawn plus, which with any imprecision on my part could soon play a role. I believe that in this variation the game is drawn with perfect play. I don't think that my move is objectively better, but it keeps my material plus, which I could hope to use in the future.)
- Freeing the knight from the pin.
- Now a discovered check is threatened and White must give back the exchange or repeat moves.
- The only way to keep the game going for a bit.
33...Bh3+ 34.Rxh3 Qxf4+ 35.Kg2 Ne5
- ( 33.Kg1 Qg5+ 34.Kf1 Qf6 repeats. )
36.Be2 d5 37.Rg3
- Attacking the bishop, and thus bringing the knight into the game with tempo -- always a good idea!
- Threatening Qxd5, which is now possible because it would threaten Qg8 checkmate, so Black wouldn't have the time to capture the undefended knight at c2.
- ( 37.Nd4 Solves the problem White has in the game, but replaces it with another problem of the same kind: now the bishop is pinned. )
- and here I offered a draw, as I can't escape perpetual check.
- ( 38.Bxd3 exd3 39.Rxd3 Qg5+ 40.Rg3 Qf4! attacking the rook, so that White doesn't play his king to the first rank, freeing his extra piece. The knight at c2 is thus as good as if it was a pawn, and the position is dead drawn. (It is never too late to lose a game! 40...Qh4?? 41.Re3 and the rook is defended by the knight, so White can answer a queen check with Kf1, freeing the knight. I had no doubt though that Rosario would have seen that. 41...Qg5+ 42.Kf1 +-) 41.Kh3 The only move to keep defending the rook AND move the king out of the pin. Problem is, it allows Black a check. 41...Qh6+ 42.Kg2 Qf4 = and the position is repeated.)
This game gave me enormous pleasure, even though I wasn't able to finish off my opponent when he played something I knew should be bad. All the same, it was great fun, as any game with Rosario, win or lose. And that is the true spirit of SchemingMind, I think -- we are all together in it for our love of chess!
I went back and looked at my notes on this game.
I was looking for White to play 17.0-0 after which I liked 17...f5.
In the actual game Kajetan plays 17.Nce3, after which I saw 17...g6 18.0-0 f5 19.Qd3 f4 20.Nc2 f3 21.g3 h5 and Black has good attacking chances on the King side.
But that was not to be as White plays 18.h4 and 18...Bxh4 seemed the only logical reply.
Would I have done anything differently ? Yes. After going over Kajetan's most excellent analysis I have to agree that 17...Bxe3 is more promising for Black.
This was a very instructional game for me and much credit to Kajetan for a fine article on it.
I always appreciate the time taken from players here to share their thoughts in such detail. It is yet another gift of the site that people who play here are civil not only in gameplay but in discussion of this game, and the enlightening observations they can make while they give over so much of their discretionary time, and with such passion. I will have to print out the piece in detail, but so as not to forget to post my initial thoughts, I am sitting here (with cup of tea at hand) enjoying what you posted, Kajetan. Many thanks, and good games to you.