Sunday, 31 May 2015

A couple of recent games in the Sicilian Najdorf

I've had a significant operation recently and been recovering, so haven't had much time to concentrate on chess (though I did get the coverage of the King's Gambit completed earlier).

I have a couple of deeply annotated games in the Sicilian Najdorf with 1.e4 c5 2.d4 cxd4 3.Nf3 d6 4.Nxd4 Nf6 5.Nc3 a6 6.Bg5 that I played recently.

Note the fairly unusual move-order.  This move-order probably won't be right for most players: it works well if you're happy to play the White side of the Morra Gambit or the ...d6 lines of the Open Sicilian (while side-stepping lines like the Kan, Taimanov and Lowenthal).  The main issue with it is that after 3...a6, 4.c3 is probably best, leading to a Morra, since after 4.Nxd4, after a subsequent ...e7-e5, White doesn't have the b5-square available for the knight on d4, and that after 3...d6 (as played in these games), 4.c3 Nf6 is awkward since 5.e5 (probably best) is now met by 5...dxe5, rather than a transposition into normal c3-Sicilian lines.

These two games highlight the downside of getting involved in these sort of highly theoretical tactical lines- at club level most of us don't really know what we're doing!  But they do tend to produce pretty interesting games.

Of course, 6.Bg5 is one of the most "theoretical" responses to the Najdorf and there are plenty of ways of playing the Open Sicilian with White that are not as theoretical, though constructing a full Open Sicilian repertoire that both avoids heavy theory and maintains good attacking chances is quite tricky- some compromises will be needed one way or another against certain lines.  For those who are interested, Michael Goeller at his Kenilworthian blog, back in January 2010, suggested a relatively aggressive and low-theory approach, mostly involving early f2-f4 advances, but also see the comments section at the end of the article for an illustration of the challenges involved.


  1. After criticizing Goeller's article I wondered if I could to a better job, but haven't exactly succeeded. For instance Van der Wiel's 1.e4 c5 2.Nf3 e6 3.d4 cxd4 4.Nxd4 Nc6 5.Nc3 a6 6.Bf4 is funny. But 4...e6 5.Nc3 b5 6.Bd3 (Goeller refers to a Kavalek article for this variation) Qb6 scores very well for Black.
    Also the Accelerated Dragon remains a hard nut to crack if White doesn't want to play the main main lines. 1.e4 c5 2.Nf3 Nc6 3.d4 cxd4 4.Nxd4 g6 5.c4 Nf6 6.Nc3 Nxd4 7.Qxd4 d6 8.Bg5 Bg7 9.Qd2 O-O 10.Bd3!? looks like an easy and strong option (plan O-O, Re1, Nd5). But White must play something different after 6...d6. The Montevideo Attack is funny (5.Nc3 Bg7 6.Nb3 Nf6 7.g4) but can be avoided with 6...d6. Taylor suggests to play the Alekhine Attack 5.Nc3 Bg7 6.Nb3 Nf6 7.Be2 d6 8.Be3 O-O 9.f4 Be6 10.g4 but 9...a5 10.a4 Be6 11.O-O forces a positional game.
    All the best with your recovery.

  2. So recently I wondered how dubious the O'Kelly really is for Black.
    1.e4 c5 2.Nf3 a6
    a) 3.d4?! cxd4 4.Nxd4 e5 as you mentioned.
    b) 3.Nc3 e6 4.d4 cxd4 5.Nxd4 b5 6.Bd3 Qb6 is a line I mentioned above.
    c) 3.c4 e6 4.d4 cxd4 5.Nxd4 Nf6 6.Nc3 Qc7 followed by ...b6 and ...Bb7 does surprisingly well for Black.

    These lines seem to show that after 1.e4 c5 2.Nf3 e6 3.d4 cxd4 4.Nxd4 a6 the only good move 5.Bd3. That's quite depressing if you want to cut down on theory for White.

    d) 3.c3 e6 4.d4 d5 5.e5 (5.exd5 exd5 is similar to the French Tarrasch; White's c3 might be premature and Black's ...a6 is useful in any case) Bd7 6.Be2/6.Bd3 cxd4 (Bd7 7.Bxb5+ axb5 8.dxc5 Bxc5 9.b4 Bb6 10.Na3 wins a pawn) 7.cxd4 Bb5 is reasonable for Black. But the clever waiting move 6.a3 is annoying.

    1. In the Kan I like the look of 1.e4 c5 2.Nf3 e6 3.d4 cxd4 4.Nxd4 a6 5.Nc3 Qc7 6.Qf3, but as you mention via transposition, after 5...b5 it is much harder to see a good alternative to 6.Bd3 (and then 6...Qb6 indeed; of course if 6.Qf3 b4). Your opinion is matched by the database at 5.Nc3 scores just 48% for White, 5.Bd3 a respectable 54%.

      As White I'd go 1.e4 c5 2.Nf3 a6 3.d4 cxd4 4.c3 but as you've pointed out elsewhere, the Morra isn't the greatest way of cutting down on theory either. 3.c3 is the best-scoring reply in the database. I'd probably slightly rather be White after 3...e6 4.d4 d5 5.e5 Bd7 6.Bd3 cxd4 7.cxd4 Bb5 8.Bc2 Bb4+ 9.Bd2 Bxd2+ 10.Qxd2. White is slightly ahead in development and has possibilities of a kingside attack with the c2-bishop pointing at h7 and the b5-bishop unable to get back to the kingside easily.

    2. "6.Bd3 cxd4 7.cxd4 Bb5 8.Bc2 Bb4+"
      is a highly depressing line imo. But I found 8...a5 9.Nc3 Ba6 interesting. It's terribly slow of course, but with a closed centre and the inability to castle quickly it will be hard to take benefit.

  3. I found some consequential improvements on the lines in your pierce gambit page but there's no place to comment there.

    1. IS is OK with you posting those improvements here. Tip: mention the games you are writing about. That makes it easy to look things up.
      It can take quite a while before he answers, but he always does.

    2. Indeed, you can post the improvements here. I'd be interested to see them.

  4. The easiest way in terms of study time to meet the Dragon proper and the Accelerated Dragon remains the Classical with Bg5. If White insists on an attacking game there are some funny shortcuts in the 9.Bc4 Yugoslav Attack.

    Soltis: 9.Bc4 Bd7 10.h4 Ne5 11.Bb3 h5 12.O-O-O Rc8 13.Rdg1 (played by GM De Firmian) Nc4 14.Bxc4 Rxc4 15.g4 (eg Hakki-Tolnai, 1982) is the same as 11...Rc8 12.O-O-O Nc4 13.Bxc4 Rxc4 14.g4 h5 15.Rdg1 (iso the usual 15.gxh5).

    Two other main lines: 9.Bc4 Bd7 10.h4 Ne5 11.Bb3 Rc8 12.O-O-O Nc4 13.Bxc4 Rxc4 14.g4 (iso the usual 14.h5) Qa5 15.h5 (or 15.Nb3 and 16.h5) and 11... Qa5 12.O-O-O Rfc8 13.g4 Nc4 (there is ...b5 of course) 14.Bxc4 Rxc4 15.h5 are nearly the same, whether Black will sac the exchange on c3 or not.

    This allows White also to play 1.e4 c5 2.Nf3 Nc6 3.d4 cxd4 4.Nxd4 g6 5.Nc3 Bg7 6.Be3 Nf6 7.Bc4 O-O (Qa5 8.O-O O-O 9.Nb3 Qc7 10.Re1!? is simple and interesting - the plan is Bg5 and Nd5) 8.Bb3 d6 (there is ...a5 of course) 9.f3 Bbd7 and as 10.Qd2 Nxd4 11.Bxd4 b5 is not easy to meet (I've played this as Black myself and enjoyed the active counterplay) I suggest 10.h4, which easily can transpose to the lines above.
    Of course Black has more options.

    Finally this allows White also to play 1.e4 c5 2.Nf3 Nc6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 d6 6.Bc4 Bd7 7.Bb3 g6 8.f3 etc.