Home > Fun Games, GM Repertoire > Testing a Negi Recommendation

Testing a Negi Recommendation

Andrew Greet (2454) – Adam Hunt (2437)
4NCL 10.01.2015

Adam Hunt has always been a tough opponent for me. I lost to him several times as a kid, drew some games here and there, but had never beaten him. Our most recent clashes were in 2010, when I drew one game from a winning position and lost another in which he simply outplayed me. In short, I felt like I needed to do something different to break out of the cycle of bad results. It just so happens that Adam is a lifelong Najdorf player and, at the time when the game took place, I had not long finished editing Parimarjan Negi’s superb 1.e4 vs The Sicilian I against this very opening. Despite not having played against a Najdorf in well over a decade, and never having played 6.Bg5 in a serious game in my entire life, I decided this would be a good moment to roll the dice. To make matters more interesting, the book was not yet published and I didn’t have any of it saved on my laptop, so my preparation was based entirely on my memory of editing the book. Fortunately Parimarjan did an excellent job of explaining the most important ideas, so I felt like I would have reasonable chances to bluff my way through any unfamiliar territory.

1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 a6 6.Bg5 e6 7.f4 Be7 8.Qf3 Qc7 9.0–0–0 Nbd7

My pre-game prep revealed that Adam had tried a few different lines against 6.Bg5, but the Three-Piece System (to use Negi’s terminology) had been his most frequent choice.


Negi mainly focuses on the main line of 10.g4, in which he also has some excellent ideas for White. However, I decided to go for the text move, which is covered as a secondary option. I chose it partly for surprise value, and also because I was familiar with the main plans and knew I wouldn’t have to recall too many complicated variations.


This is Black’s most popular reply, and Adam did not spend much time on it.

10…Ne5 is another move, when Negi recommends 11.Qh3 to hit e6, rather than 11.Qg3 which he himself once played.


11.Nb3 has been more popular, but it will soon become clear that the knight is better on e2.

11…b5 12.Bxf6 Nxf6 13.Nd5 Nxd5 14.exd5 Bb7N

The first new move, although it is an obvious choice which Negi gave as his main line.

14…b4 has been played before, but after 15.Ng3 Black followed up with 15…Bb7 anyway in Grabarczyk – Matuszewski, Grodzisk Mazowiecki 2007. Hunt’s move is more accurate, as Black has no reason to hurry with …b4.

15.Ng3 Rc8 16.Bd3 Qc5 17.Be4

I couldn’t remember if this exact position was in Negi’s book, but I knew that this was more or less White’s ideal set-up. The bishop on e4 does a perfect job of guarding the weak points on c2 and d5. The knight is poised to go to h5, followed by a kingside advance. Black has no obvious counterplay, his bishops have limited scope and his king faces an uncertain future. In short, the outcome of this rare opening line  is excellent for White, so this can be considered a superb recommendation from Parimarjan.

17…Bg5† 18.Kb1 Qe3 19.Qh5

Although I was unaware of it during the game, this is actually the point where Negi ends his analysis. He gives a ‘plus-equals’ symbol with the comment: “White will soon drive the enemy pieces away with Rd3 and h2-h4.”

19…Bh6 20.Rd3 Qg5 21.Qe2 0–0 22.h4

I spent more than half an hour considering the various possibilities here. The key attacking idea is f5-f6, but should I play it immediately, or after h2-h4, or should I put a rook on the f-file first, and if so, which rook?



23.Nh5 f6 24.g4 is another possibility, but I thought I risked losing control after a timely …Bf4, when exchanging on f4 might win a pawn but allow counterplay along the e-file.

23…Qxf6 24.Rf1 Bf4

24…Qd8 25.Nf5 is unpleasant for Black.

25.Nh5 Qd8?

25…Qh6 is better, although 26.g4 (26.g3? f5!) 26…f6 27.Bf5 gives White an excellent position.


I overlooked 26.Nxg7!, with the point that 26…Kxg7 27.Qg4† is crushing.


Now Black is better, although White is very much in the game. Obviously we are past the point of relevance to the opening. Still, the rest of the game contains some interesting moments.

27.gxf4 fxe4 28.Rg3 Rc7 29.Rfg1 g6

29…Rff7! 30.Qxe4 Qf8! is the more precise continuation proposed by the metal box.

30.Qxe4 Kh8 31.fxe5 gxh5

31…Rcf7 32.Qe1! gxh5 33.e6 forces the rook back to c7.


During the game I thought this looked extremely dangerous for Black. The position is in fact dynamically equal.


32…Qf6 33.Qg2 Rcc8 34.a3 leaves Black tied up, but it is hard to improve White’s position, so I think it should be a draw.

33.Qe3 Qe7 34.b3

34.Rg5! would have put Black in trouble.

34…Rg4 35.Rxg4 hxg4 36.Rxg4?

Better was 36.Qd4† Qg7 37.Qxg7† Kxg7 38.Rxg4† Kf6 39.Rf4† Ke7 40.Rxf8 Kxf8 41.c4 with a drawn endgame.


36…Rf1† 37.Kb2 Bxd5 38.Qd4† Qf6 39.Qxf6† Rxf6 40.Rd4 Bxe6 41.Rxd6 The game goes on, but White is struggling.

37.Rd4 a5 38.a3 b4 39.Ka2 Rc8 40.Rc4


Allowing the rooks to come off is an ill-fated decision for Black.

40…Rg8! would have maintained an edge for him.

41.Rxc8† Bxc8 42.c3 Qe7 43.Kxa3 Kg7 44.Qd4† Kg8 45.Qc4 Qb7

45…Bxe6 does not work due to 46.Qe4.

46.Ka4 Kf8 47.Qf4† Ke8 48.Qf6

It would have been lovely to have won a perfect attacking game (26.Nxg7!), but things rarely go so smoothly when facing a tough opponent. Still, I enjoyed testing Negi’s repertoire over the board, and if a ‘6.Bg5 virgin’ like me can use it to get a clear advantage, there must be something to it. 1.e4 vs The Sicilian I by Parimarjan Negi is published on 4th February 2015.

Categories: Fun Games, GM Repertoire Tags:
  1. TonyRo
    February 3rd, 2015 at 23:00 | #1

    Nice game Andrew, and a good advertisement for the book. After 20 moves or so Black’s position looked pretty prospectless…

  2. Jacob Aagaard
    February 4th, 2015 at 09:10 | #2

    I have to second that. Great prep and great game.

  3. Jesse
    February 4th, 2015 at 14:07 | #3

    Nice win Andrew.

    Beating a life-long rival for the first time is a major accomplishment, even more elusive than GM norm.

  4. Johnnyboy
    May 14th, 2016 at 17:58 | #4

    Noticed Sicilian 1 is missing an important line that is actually mentioned as the best move in the particular position in Experts by Luther. On p215 in the 3 piece system line B2211 14..Ne5 15. Qh5 the immediate 15….Bf6 is not mentioned…only after the inclusion ….Qd8 h4. Both taking with the pawn and Knight on e6 look viable. Does Parimarjan have a recommendation?

  5. Nikos Ntirlis
    May 15th, 2016 at 11:46 | #5

    Corr players seem to believe that allowing Nxe6 is very unpleasant, that’s why the move 15…Qd8 is more popular (in that case Nxe6 is not good as Negi suggests). Everything else (15…Qe7 -given as an example by Negi- , 15…Bf6 or even 15…h6 -unmentioned also by Negi-) allow the very good 16.Nxe6! with a safe advantage for White in all cases. This advantage might be big or small, i cannot really tell for sure, but from a practical perspective i like White in both OTB and corr play.

    I remember some old lectures i attended by Marjanovic, where he told us that from Black’s point of vies, the most important task is to take the f-file under control, otherwise he is lost. Something like …Ra7-Rf8 to f3 and Rf7 was the only viable Black’s plan. Even then though, the modern engines preffer White’s chances. This is maybe the reason in practice that this line has been abandoned by Black today.

  6. Jacob Aagaard
    May 15th, 2016 at 14:27 | #6

    1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Bg5 e6 7. f4 Be7 8. Qf3 Qc7 9. O-O-O Nbd7 10. g4 b5 11. Bxf6 Nxf6 12. g5 Nd7 13. f5 Bxg5+ 14. Kb1 Ne5 15. Qh5 h6 16. Nxe6 Bxe6 17. fxe6 g6 18. exf7+ Qxf7 19. Qh3 Rd8 20. Nd5 O-O has been played in four correspondence games. Basically, White has a big advantage. Strongest is possibly: 21. Qa3 and White is better. Computers might be able to hold this against computers in correspondence chess, but this is true for many bad positions. I would be happy to play White.

  7. Johnnyboy
    May 15th, 2016 at 17:55 | #7

    Thanks Nikos
    In the position after 15…Bf6 after taking your advice to take with the knight, 16.Nxe6 Bxe6 fxe6 O-O Nd5 Qd8 seems best but after that (with my limited analytic skills) exf7, Bh3 and even Nxf6 and then capture the d6 pawn. In all 3 lines black has an unweakened pair of pawns protecting the king unlike lines such as Jacob’s where the weakening …h6 have been played. What choice was made in those corr games or do you recommend? Looks defendable to me

  8. Johnnyboy
    May 15th, 2016 at 18:17 | #8

    In a related matter where do you find these correspondence games? I know TWIC covers over the board games but you book and Negi’s have drawn heavily on corr games as the leading edge of theory rather than otb games. Thanks

  9. Jacob Aagaard
    May 16th, 2016 at 07:07 | #9

    Nikos spends a lot of time on the Internet going to various places finding games.

  10. Johnnyboy
    May 16th, 2016 at 10:11 | #10

    And are you willing to share these various places or do I need a secret QC handshake to be in the inner circle?

  11. Jacob Aagaard
    May 16th, 2016 at 10:25 | #11

    I don’t know them to be honest.

  12. Johnnyboy
    May 16th, 2016 at 11:05 | #12

    same here!

  13. May 16th, 2016 at 20:46 | #13

    Start with ChessBase’s 2016 Correspondence Database. I reviewed it on my website. You need to keep the database up to date, but luckily this is fairly easy.

    ICCF is the largest source of new high quality correspondence games. You have to sign up with the website to download the monthly archives.

    Other sources:
    BdF: http://www.bdf-schachserver.de/
    LSS: http://www.chess-server.net/
    DESC: http://www.desc-online.de/service/archiv/download.php

  14. Johnnyboy
    May 17th, 2016 at 14:34 | #14


    many thanks for the weblinks. Another question though-How can you tell the quality of corr games- the casual corr game from the top level ones? Can use ELO for otb games but not sure how best to evaluate corr games quality.


  15. Björn
    May 17th, 2016 at 14:50 | #15

    Johnnyboy: Elo is not such a bad guide either for corr games. However, looking only at Elo > 2400 or so, as one might do for otb games is probably too restrictive, because even at lower Elos people will play pretty well in corr games. This is because engines are allowed and used in most competitions. Thus, I tend to take corr games by 2100 or 2200 rated players pretty seriously (and even at lower ratings you get the occasional gems, where some minor opening variation is refuted etc.). Another hint might be the competition, e.g. the finals, semi-finals or quarter-finals of the world championship, or Olympiads will definitely be fiercely competed.

  16. May 17th, 2016 at 14:59 | #16

    What Bjorn said. I actually filter out games where both players are 2000+ and make a Chessbase opening tree out of it each month, using the result to check quickly for high scoring moves, etc.

  17. Johnnyboy
    October 2nd, 2016 at 16:14 | #17

    Jacob Was Boris not given a copy of Negi Sicilian 1? Down an hour on the clock today and still in Negis recommendation. Thought what Boris doesn’t know about the Naylor isn’t worth knowing while Kramnik who is a rare Sicilian player as white had more time left than when he started when he finished Boris off. His longest think was 11 minutes after following Parimarjan recommendation

  18. Johnnyboy
    October 2nd, 2016 at 16:15 | #18

    Naylor = Najdorf

  19. Remco G
    October 3rd, 2016 at 07:27 | #19

    @Johnnyboy: and when Giri beat Gelfand, he said “Boris really loves to sacrifice pawns, so in the game I was happy that I ended up a pawn down. Before the tournament I read Gelfand’s book, so I understood how he evaluates.”

  20. Johnnyboy
    October 3rd, 2016 at 11:38 | #20

    I’m sure someone else (was it Giri?) had said he read Boris’s book so knew how he evaluates as well and that helped him choose what to play. maybe the book was good for others and not so good for Boris himself.

  21. Jacob Aagaard
    October 3rd, 2016 at 11:42 | #21

    Don’t tell Boris!!

  22. Johnnyboy
    October 3rd, 2016 at 11:43 | #22

    apologies it was indeed Giri who said that. Be interesting what Kramnik said

  1. No trackbacks yet.

 Limit your comments to