(c) 2000 by Sören Jensen

Sören Jensen

Nimzovich Defence with 3...dxe4, 7.Sh3!?

Nimzowitsch-Verteidigung mit 3...d:e4, 7.Sh3!?

mailto: soren@ucrac1.ucr.edu (Sören Jensen)

In this letter I will present a brief overview of one of the more critical lines in the Nimzowitsch defence. My aim has been to examine most games available to me and present some critical evaluation. These evaluations as well as the original analysis presented should of course be treated with great suspicion. My main sources have been: Pickard, S. & Myers, H.E., The Nimzovich Defense Ultimate CD (Pickard, 1999), Myers, H.E., The Nimzovich Defense to 1.e4 (Caissa, 1995 and earlier editions), Keilhack, R. & Schlenker, R., 1...Sc6 aus allen Lagen (Kania, 1995).

The position in question is normally reached through the move order

1.e4 Nc6 2.d4 d5 3.Nc3 dxe4 4.d5 Ne5 5.Bf4 Ng6 6.Bg3 f5 7.Nh3

For the pawn White has a lead in development, two strong bishops (usually positioned on g3 and c4) and with a timely f2-f3 can open up the position. Black on the other hand has plenty of weak squares, difficulty developing his king-side, and there is no obvious safe retreat for his king.

With 6...f5 Black threatens to trap White’s bishop. It is, however, not altogether clear how dangerous this threat is as Black can hardly afford to loose any more tempos. The strength of 7.Nh3 is not so much that it halts Black’s f-pawn, but that it is a natural developing move which does not require an immediate response.

The strength of this plan may in fact have dissuaded some from playing 3...dxe4, instead trying 3...e6, 3...Nf6 or 3...e5, or perhaps, even steering clear of the Nimzowitsch defence altogether. Another symptom could be the recent recommendations in some books of playing 1.e4 Nc6 2.d4 e5. I feel that 3...dxe4 is very much at the heart of the Nimzowitsch defence, though anyone who plays this regularly must be prepared to face several very critical variations. As for 1.e4 Nc6 2.d4 e5 this is of course perfectly playable but I would echo Vlassov’s sentiment on 1.e4 Nc6 2.Nf3 e5 - it's too banal.

Returning to the starting position, Black must make an important choice, opening the position with 7...e5, A) or biding time with 7...a6, B).

A. 7.Nh3 e5 8.dxe6ep

The only testing move after which Black can choose from 8...Qxd1 9.Rxd1 Bb4, A1) or 8...Bxe6, A2).


A1) 7.Nh3 e5 8.dxe6 Qxd1 9.Rxd1 Bb4

Black can also play 9...c6, which after 10.Bc4 Bb4 transposes to the text line. As discussed below there could be some merit in this move order as it avoids the line 9...Bb4 10.Bxc7, which may be a problem. After 9...c6, White’s best probably is 10.Bc4 though 10.f3 also comes into consideration.
After 9...c6 10. Bc4 Black should return to the text line with 10...Bb4. Black did not do well after 10.Bc4 h6 in Apicella-Soeteway, Brussels 1993, a game that nicely illustrates the strength of White’s bishops and a Black king trapped in the centre, 11.0-0 Nf6 12.f3 Bc5 13.Kh1 b5 14.Bb3 e3 15.Rfe1 Ke7 16.Nf4 Nxf4 17.Bxf4 Ba6 18.Ne2 Rad8 19.Rxd8 Kxd8 20.Be5 Ke7 21.Nd4 Bxd4 22.Bxd4 Nd5 23.Bxg7 Rh7 24.Bd4 Kxe6 25.Rxe3 Kd6 26.Re8 Rd7 27.Rh8 Kc7 28.Rxh6 Nf4 29.c3 Ne2 30.Bxa7 Nc1 31.h4 Rd2 32.Rh7 Kd6 33.Bb8 Kc5 34.Be6 Rxb2 35.Ba7 Kd6 36.Bxf5 Rxa2 37.Bd4 Ra1 38.Rd7, 1-0.


At first it would appear that 10.Bxc7 would help Black as he can pick up the pawn on e6 and place a rook on c8. The game Plaskett-Simons, Jersey 1998, suggests that 10.Bxc7 is far from innocuous. This game continued 10...Bxe6 11.Bb5 Ke7 12.Ng5 Nf6 13.Nxe6 Kxe6 14.Rd4 Rhc8 15.Rxb4 Rxc7 16.0-0 Ne5 17.f3 exf3 18.gxf3 Nd5 19.Nxd5 Kxd5 20.c3 a6 21.Be2 Nc6 22.Rd1 Ke5 23.f4 Kf6 24.Rb6 Re8 25.Bf3 Re6 26.Kf2 h6 27.h4 g5 28.Bd5 Rd6 29.Bxc6 Rdxc6 30.Rd6 Rxd6 31.Rxd6 Kg7 32.hxg5 hxg5 33.fxg5 Rc4 34.Kf3 Rg4 35.Rb6, 1-0. Simons enter this line one year later but I am not sure what improvement he had found. Perhaps 12...a6 should be tried, for example 13.Ba4 (13.Nxe6 axb5) b5 14.Nxe6 (14.Bb3 Bxb3 15.axb3 Rc8) bxa4 15.Nxg7 Kf6 16.Nh5 Kg5 17.Ng3 Rc8 is unclear. Perhaps even the brazen 12...Bxa2 13.0-0 Bxc3 14.bxc3 Kf6 warrants a closer look. Black can avoid this line by playing 9...c6 (see above).

10 . . . c6

In Heine Nielsen-Furhoff, Copenhagen 1991, Black played 10...Ke7 but his king proved a liability after 11.0-0 Bxc3 12.bxc3 Bxe6 13.Bxe6 Kxe6 14.f3 h6 15.fxe4 fxe4 16.Nf2 Nf6 17.Nxe4 Nxe4 18.Rfe1 Rad8 19.Rxe4 Kf5 20.Rde1 Rd7 21.h4 h5 22.R4e3 Nf8 23.Rf3 Kg6 24.Re8 Rf7 25.Rxf7 Kxf7 26.Rc8 Rh6 27.Rxc7 Ke8 28.Rxb7 Rc6 29.Rxg7 Rxc3 30.Bd6 Nd7 31.Re7 Kd8 32.Rh7 Ke8 33.Rxh5 Rxc2 34.Ra5 Nf6 35.Kh2 Rc3 36.Ra4 Nd5 37.h5 Kd7 38.Be5, 1-0.


After 11.Bc7 Black can play 11...Bxc3 (Myers played 11...b5 against a computer with the continuation 12.Rd8 Ke7 13. Bb3 Nf6 14.Rd4 Bc5 15.Rd2 Bxe6 16.Bxe6 Kex6 17.0-0 h6. The rook manoeuvre seemed like waste of time and 12.Bb3 would be stronger with an unclear position) 12.bxc3 Nf6 (queen-side expansion did not work well in Gottlieb-Six, IECG 1997, after 12...b5 13.Bb3 a5 14.a4 Nf6 15.Ng5 Ra7 16.Bb6 Ra6 17.Bc5 Nd5 18.axb5, 1-0), and now try the plan of Nd5 as in the text line, for example 13.Ng5 Nd5 14.Bxd5 cxd5 16.Rxd5 h6. Also to be considered 11.f3.

11 . . . Bxc3
12.bxc3 Nf6

Also possible 12...N8e7, likely transposing, but which also may lead to unique variations. In Wittmann-Leeners, corr. 1986, play continued 13.a4 Nd5 14.Rxd5 cxd5 15.Bxd5 Ke7 16.Rb1 Rd8 17.c4 Bxe6 18.Rxb7 Kf6, 0-1. 13. a4 looked like a waste of time.


Inferior would be 13.a4, transposing to Wittmann-Leeners above after 13...Nd5.

13 . . . Nd5


Probably also critical is 14. fxe4. If now 14...Ne3 15.exf5 Nxc4 (15...Nxf1 16.Rxf1 Ne7 and with two pawns for the exchange and the bishop pair, White should be at least +/-) 16.Ng5 Ne3 (16...Nge5 17.Rde1; 16...0-0 17.fxg6) 17.exg6 Nxf1 (17...h6 18.Nf7 Nxd1 19.Nxh8 Nxc3 20.Bd6 +-) 18.Rxf1 h6 19.Nf7 Rf8 (19...Rg8 20.Re1) 20.Re1 and Black must return material to avoid Bh4 followed by Nd6.

14 . . . cxd5

This has been played twice, but is it White’s strongest move in this position?

15 . . . dxc4
16.fxe4 f4

Instead 16...fxe4 was better for White in Mevel-Marks, corr. 1986, after 17.Ng5 Rf8 18.Rf7 Bxe6 19.Rxg7 Bf5 20.Nxh7 Rh8 21.Nf6 Kf8 22.Rxb7, 1-0

17.Nxf4 Nxf4

With complicated play in Tidman-Simons, Jersey 1999, which had the fascinating continuation, 18.Rxf4 Bxe6 19.Rd6 Ke7 20.e5 h6 21.Bh4 g5 22.Rf6 gxh4 23.Rdxe6 Kd7 24.Rxh6 h3 25.g4 Rxh6 26.Rxh6 Rg8 27.Rd6 Ke7 28.Rd4 b5 29.Kf2 Ke6 30.Re4 Rd8 31.Re2 Rd5 32.Kg3 Rxe5 33.Rxe5 Kxe5 34.Kxh3 a5 35.Kg3 b4 36.Kf3 a4 37.Ke3 b3 38.cxb3 cxb3 39.axb3 a3, 0-1. It is difficult to believe that White should have to loose from the position around move 24. White can improve with 23.Rfxe6 Kf7 24.Rxh6 but perhaps Rd8 gives chances. Perhaps better 25.Rd6 Ke7 26.Rhe6 Kf7 27.g3 and White seems better.

Conclusion: This line appears to be in danger in view of likely improvements to the game Tidman-Simons at about move 23-25, and possibly also from the untested 14.fxe4.


A2) 7.Nh3 e5 8.dxe6 Bxe6

9.Nb5 Bd6

This line may have been first recommended by Myers (1985), as an improvement over 9...Bb4, which had caused several Black disasters.


Also dangerous 10.Nxd6. Bauza-Dunne, corr. 1990, continued 10...cxd6 11.Bxd6 (11.Qxd6 Qxd6 12.Bxd6 0-0-0 13.0-0-0 Nf6 14.Nf4 Rxd6 15.Rxd6 Nxf4 16.g3 Kc7 17.Rd2 Ng6 18.Bh3 Ne5 19.Rhd1 Nf3 20.Rd6 Re8 21.Bf1 Ng4 22.Bb5 Re7, 0-1, Jensen,B.-Hill corr. 1996) 11...Qb6 12.Ba3 Rd8 13.Qe2 a6 14.Rd1 N8e7 15.Rxd8 Qxd8 16.Qd2 0-0 17.Ng5 Bd5 18.h4 Re8 19.h5 e3 20.fxe3 Ne5 21.c4 f4 22.e4 h6 23.Nf3 Nxc4 24.Bxc4 Bxc4 25.Qxd8 Rxd8 26.b3 Bb5 27.Bxe7 Re8 28.a4 Bc6 29.Bc5 Rxe4 30.Kf2 b6 31.Bd6 Re3 32.Nd4 Rd3 33.Nxc6 Rxd6 34.Rc1 Rd2 35.Kg1, 1-0. Possibly better 13...N8e7, 14.Qb4 Qxb4 15.Bb5 Bd7 16.Bc4 h6, or 11...Qa5 12.b4 Qb6 13.Bc5 Qc6.

10. . . cxd6

This has been popular following the game Rogers-Dunne. But even more critical may be 11.Qxd6 since 11...Qxd6 12.Nxd6 Ke7 13.Nxb7 (diagram)
and 13...Rb8 does not return the pawn, as pointed out by Keilhack & Schlenker. In Nabours- Dunne, corr. 1992, Black instead played 13...Rc8 but did not equalize after 14.0-0-0 Bxa2 15.Nd6 Rc5 16.b4 Re5 17.Nc8 Kf6 18.c4 e3 19.Rd6 Kf7 20.fxe3 Rxe3 21.Kb2 Bxc4 22.Bxc4, 1-0.

11 . . . Nf6

Here 11...Nh6 was no improvement in Rogers-Laird, Sun Coast 1998, after 12.Nxd6 Kf8 13.0-0-0 Nf7 14.Nxf5 Qxd4 15.Nxd4 Bxh3 16.gxh3 Nf4 17.Rg1 Rc8 18.Rg3 g5 19.Rb3 b6 20.Ra3 a5 21.Rb3 Kg7 22.Rxb6 Rc5 23.Ra6 Rd8 24.Nb3 Rxd1 25.Kxd1 Rd5 26.Kc1 h6 27.Rxa5 Rd6 28.Nd2 Re6 29.Nc4 Rc6 30.Ra7 Ng6 31.Nd2, 1-0.

12.Ng5 Bd5
13.Bc4 Ne7

Critical. 13...Nf4 did not work in Rogers-Dunne, Philadelphia 1987, 14.g3 Rc8 15.b3 a6 16.gxf4 axb5 17.Bxd5 Rxc2 18.0-0 Qa5 19.b4 Qa3 20.Bb3 Rc6 21.Nf7 0-0 22.Ne5 d5 23.Bxd5 Kh8 24.Bb3 Rc7 25.Qd6 Rcc8 26.Rac1 Qxc1 27.Qxf8 Rxf8 28.Rxc1 Nh5 29.Nf7 Kg8 30.Ng5 Kh8 31.Ne6, 1-0
There is some difference in opinion over 13...Bxc4. Rogers gave 14.Nxd6 Kf8 15.Qxc4+/-. Analysis in the Correspondence Chess Yearbook 11 continued 15...Qa5 16.c3 Ne5 17.Qb3 Nd3 18.Kf1 Qd5 with decent play for Black. White can improve with 16.Kf1.


14.Ne6 led to a complicated game in Hart-Dunne, corr. 1991, 14...Bxe6 15.Nxd6 Qxd6 16.Qxd6 Bxc4 17.0-0-0 Kf7 18.Rd4 Rac8 19.Kb1 Rc6 20.Qb4 Rhc8 21.Qxb7 Bd5 22.Qxa7 Rxc2 23.Rhd1 e3 24.fxe3 Be4 25.Rxe4 Nxe4 26.a4 Nd2 27.Ka1 Kf8 28.a5 Rd8 29.Ka2 Nc6 30.Qb7 Rd5 31.Rxd2 Rdxd2 32.Qc8 Ke7 33.Qxf5 Rxb2 34.Ka3 Ra2 35.Kb3 Nxa5 36.Kb4 Rdb2 37.Kc5 Nb3 38.Kc6 Rc2 39.Kb6 Kd6 40.Qf4 Kd5 41.Qf5 Kc4 42.Qe4 Kc3 43.Qe5 Kd2 44.Qxg7 Nc1 45.e4 Rab2 46.Ka6 Nd3 47.Qg5 Kd1 48.Qe3 Rb3 0-1

14. . . a6

The somewhat mysterious 14...Qb8 was suggested in CCYB11. After 15.Ne6 Bxe6 16.Nxd6 Qxd6 17.Qxd6 Bxc4 Black is in fact a tempo down on the text-line. Perhaps the point is that the square a6 is now available as a retreat square for the bishop.

15.Ne6 Bxe6
16.Nxd6 Qxd6
17.Qxd6 Bxc4

With a delightfully unbalanced position in Glaser-Dunne, corr. 1993, 18.b3 Bd5 19.c4 Bc6 20.Rhe1 Kf7 21.f3 Rhc8 22.g4 Rd8 (22...fxg4 23.fxe4 Bxe4) 23.Qe5 Rxd1 24.Kxd1 Rd8 25.Kc2 fxg4 (25...g6 26.gxf5 gxf5 27.Rg1 Rd7 28.fxe4 Bxe4 29.Kc1 Nc6) 26.fxg4 Ng6 27.Qa5 Re8 28.Qc7 Re7 29.Qg3 h6 30.Re2 Nf8 (30...Rd7 31.h4 Rd3 32.Qc7 Rd7 33.Qg3 Rd3=) 31.g5 hxg5 32.Qxg5 Ne6 33.Qe5 Rd7 34.Rd2 Rxd2 35.Kxd2 g6 36.b4 Nd7 37.Qc3 Ng5 (37...Nf6!?) 38.b5 axb5 39.cxb5 Bd5 40.a4 Nf3 41.Ke2 Nde5 42.Qc5 Bc4 43.Ke3 Bd3 44.Qc7 Kf6 45.Qxb7 g5 adjudicated 1-0 (comments from CCYB 11).

Conclusion: Dunne’s spectacular sacrifice of queen for three light pieces appears to make this line playable against 11.Qd4, though this type of position may not be to everyone’s liking. A bigger problem, however, appears to be 11.Qxd6 Qxd6 12.Nxd6 for which there is no satisfactory defence.


B. 1.e4 Nc6 2.d4 d5 3.Nc3 dxe4 4.d5 Ne5 5.Bf4 Ng6 6.Bg3 f5 7.Nh3 a6


This is the critical move but 8.f3 also needs study. Dunne-Roques, IECG 1997, appears to show a useful defence 8...e5 (8...Nf6 looks inferior. Black drew in Cirello-Calua corr. 1994 after 9.fxe4 fxe4 10.Ng5 Bf5 11.Be2 Qd7 12.0-0 e5 13.dxe6 Bxe6 14.Qxd7 Bxd7 15.Bc4 Bc5 16.Kh1 Be3 17.Ngxe4 Nxe4 18.Nxe4 0-0-0 19.Rae1 Bd4 20.Ng5 Rde8 21.c3 Rxe1 22.Rxe1 Bf6 23.Ne4 Bd8, ˝-˝, but 11.Bc4 looks better) 9.dxe6 Bxe6 10.fxe4 (Myers also recommend 8...e5, giving here 10.Qxd8 Rxd8 11.Ng5 Bd7 12.Bc4 Nh6 and some additional analysis) 10...Qxd1 11.Rxd1 Bb4 12.Ng5 Bxa2 13.exf5 N6e7 14.Ne6 Bxe6 15.fxe6 Nf6 16.Bxc7 Rc8 17.Bd6 Bxc3 18.bxc3 Nf5 19.Bb4 Rc6 20.Bd3 Rxe6 21.Kd2 g6 22.Rhe1 Rxe1 23.Rxe1 Kd7 24.Bxf5 gxf5 25.Re7 Kc8 26.c4 Rg8 27.g3 Rg4 28.Kd3 b5 29.cxb5 Nd5 30.bxa6 Nxe7 31.Bxe7 Ra4 32.c4 Kd7 33.Bf8 Rxa6 34.Ke3 Ke6 35.c5 Ra4 36.Bd6 Kd5 37.Kf3 Re4, ˝-˝.
Less challenging is 8.Nf4. Kayser-Zimmer, corr. 1998, continued 8...Nxf4 9.Bxf4 Nf6 10. Bc4 b5 11.Bb3 Bb7 12.a3 Qd7 13.Qd2 c5 14.dxc6 Qxd2 15.Bxd2 Bxc6 16.Bf4 Rd8 17.0-0 Nd5 18.Ne2 e6 19.Be5 Kf7 20.Nd4 Bb7 21.c3 Bd6 22.Rae1 Bxe5, 0-1. That ended with a dropped piece but by this stage Black was doing fine.

After 8.Bc4 Black has a multitude of moves: 8...h5, will be examined in B1), and 8...b5 and other moves in B2).

B1) 7.Nh3 a6 8.Bc4 h5

Possibly new, the aim of this positionally questionable move is to ease Black’s cramped position by exchanging or displacing White’s black-squared bishop.


Less testing is 9.Nf4 which did not trouble Black in Wu-Jensen, corr. 1999, after 9...Nxf4 10.Bxf4 g6 11.0-0 Bh6 12.Qd4 Nf6 13.Be5 Rf8 14.Qc5 Nd7 15.Qxc7 Nxe5 16.Qxe5 Qd6 17.Qxd6 exd6 18.Ne2 Bg7 19.c3 Be5 20.f3 exf3 21.Rxf3 Bd7 22.Re1 Kf7 23.g3 Rfe8 24.Kf2 Rac8 25.Bd3 b5 26.Bc2 a5 27.h3 Bf6 28.Nf4 Rxe1 29.Kxe1 a4 30.Kd2 h4 31.gxh4 Bxh4 32.b3 Bf6 33.Ne6 Be5 34.Nd4 axb3 35.axb3, ˝-˝.
Also 9.f3 needs attention after which Black should perhaps play e3.

9 . . . h4
10.Bf4 Nf6

Tempting but probably inferior is 10...Nxf4. Now 11.Nxf4 Qd6 12.Ne6 Nf6 (12...Bxe6 13.dxe6 0-0-0 looks too dangerous) 14.g3 g6 15. 0-0-0 h3 (13...Bxe6 14. dxe6 Qf4 15. Kb1 worked in Whites favour in Gaard-Jensen friendly corr. 1999) 14. Bb3 Bh6 16. Kb1 Ld7 17.f3 exf3 18.Qxf3 Bxe6 (this looks dubious but Black felt forced to do something and the White pawn on e6 gives prospect of castling) 19.dxe6 Qb6 20.Nd5 Nxd5 21.Qxd5 0-0 with a slight advantage to White in Gaard-Jensen, friendly corr. 2000).
10...e5 11.dxe6ep Nxf4 12.Nxf4 Bd6 also could be considered.

11.0-0-0 Nh5

And now, for example, 12.Bg5 Qd6 13.f3 Nhf4 14.Nxf4 Nxf4 15.Qd2 e3 16.Qxe3 Nxg2 17.Qf2 h3.

Conclusion: This line need further testing, though I remain quietly optimistic about Black’s chances here.


B2) 7.Nh3 a6 8.Bc4 b5
(8...Nf6, 8...Nh6, 8...e5)

1) 8...Nf6 often transposes to positions starting 8...b5, but unique positions are also possible.

1a) 9.f3 b5 (9...e5 10.dxe6 Qxd1 Rxd1 +/-, Myers) 10.Bxb5 axb5 11.Nxb5 Kf7 12.Nxc7 Rb8 13.Ne6 Qa5 14.c3 Rxb2 15.0-0 Qxd5 16.Qxd5 Nxd5, analysis by Sorensen, in Myers 1995. It would seem more sensible for White to keep the tension with 10.Bb3
1b) 9.Qe2 e5 (9...b5 10.Bxb5 axb6 11.Nxb5 Ba6 12.Nxc7 Qxc7 13.Qxc7 Bxe2 14.Kxe2 Nxd5, Glasscoe, quoted in Myers 1995. 10. Bb3 may be sounder) 10.dxe6ep Bb4 11.Rd1 Qe7 12.0-0 Bxc3 13.Bxc3 b5 14.Bb3 0-0 is good for Black according to analysis of Boey, quoted in Myers, but 15. Bxc7 looks very unpleasant.
1c) 9.0-0. Black used an interesting plan in Ottevaere-Vandermeulen corr. 1996, 9...b5 10.Bb3 Bb7 (diagram)
11.Qe2 Qc8 12.Rad1 e5 13.dxe6 c5 14.Bd5 Bxd5 15.Nxd5 Nxd5 16.Rxd5 Qxe6 17.Rfd1 Be7 18.a4 bxa4 19.Qh5 0-0 20.Bd6 e3 21.Bxe7 e2 22.Re1 Qxd5 23.Bxf8 Re8 24.Ng5 Qd1 25.Qxh7 Kxf8 26.Nf3 Ne5 27.Qh8 Kf7 28.Ng5 Kf6, 0-1. An earlier f3 would have been more testing.
1d) 9.Bb3 was played in LeBlancq-Wood, London 1987, 9...e5 10.dxe6 Be7 11.Qe2 c6 12.Rd1 Qa5 13.0-0 0-0 14.Bd6 Kh8 15.Nf4 Bxd6 16.Nxg6 hxg6 17.Rxd6 Qe5 18.Rfd1 Ng4 19.g3 Qf6 20.f4 Re8 21.Nxe4, 1-0.
1e) Less dangerous 9.Ne2, as in Jones-Thomas, Chepstow 1999, 9...b5 10.Bb3 e5 11.dxe6 Qxd1 12.Rxd1 c5 13.Bc7 Be7 (what happens on 13...c4?) 14.c3 c4 15.Bc2 Bxe6 16.Nd4 Bd7 17.0-0 Ra7 18.Ba5 Ng4 19.Rfe1 0-0 20.f3 exf3 21.gxf3 N4e5 22.Bb4 Bxb4 23.cxb4 Nc6 24.a3 Nxd4 25.Rxd4 Bc6 26.Ng5 h6 27.Ne6 Rf6 28.Rd8 Kh7 29.Nd4 Re7 30.Rd1 Nh4 31.Kf2 g5 32.Rc8 Be8 33.Rc5 Bg6 34.Nc6 Rc7 35.Ne5 Re7 36.Nd7 Rff7 37.Ne5 Rf8 38.Nd7 Ra8 39.a4 Kg7 40.Rc7 f4 41.Bxg6 Nxg6 42.Rd6 Ne5 43.Nc5 Rxc7 44.Ne6 Kf7 45.Nxc7 Rc8 46.Nxa6 bxa4 47.Nc5 Nd3 48.Nxd3 cxd3 49.Rxd3 Rc2 50.Kg1 Rxb2 51.Rd7 Ke6 52.Ra7 Rxb4 53.Kg2 Kd5 54.Kh3 Kc4 55.Kg4 Kb3 56.Kh5 Rb6 57.h4 gxh4 58.Kxh4 a3, 0-1.

2) Myers recommends 8...Nh6 with the plan of Nf7, Nfe5. This logical plan is rather slow but worth further testing. It did not work well in Johannesson-Kamminga, corr 1993, after 9.Qh5 (less energetic play gives Black an easier game as in Jensen-Beckmann, corr. 1989/90, 9.Qe2 Nf7 10.Bb3 Nge5 11.Bxe5 Nxe5 12.f3 g6 13.fxe4 Bg7 14.0-0 Qd6 15.Nf4 0-0 16.h3 Kh8 17.exf5 Bxf5 18.Ne4 Bxe4 19.Qxe4 Nd7 20.c3 Nc5 21.Qe3 Nxb3 22.axb3 Bh6 23.g3 Bxf4 24.Rxf4 Rxf4 25.gxf4 Rf8 26.Re1 Rxf4 27.Qxe7 Qxd5 28.Qxc7, ˝-˝). 9...Nf7 (the plan 9...e5 10.dxe6ep Qd4 did not work in Burk-Wrinn, corr 1989, 9...e5 10.dxe6 Qd4 11.Bb3 Be7 12.Rd1 Qc5 13.0-0 0-0 14.Nxe4 Qc6 15.Nc3 f4 16.Nxf4 Nxf4 17.Bxf4 Rxf4 18.Nd5, 1-0) 10.0-0-0 Nge5 11.d6 g6 12.dxc7 Qxc7 13.Bxf7 Kxf7 14.Qe2 Bh6 15.Kb1 Be6 16.f3 Bc4 17.Qe1 g5 18.Nd5 Bxd5 19.Rxd5 f4 20.Qxe4 Nd7 21.Re1 Nf6 22.Qe6 Kf8 23.Bf2 Nxd5 24.Qxh6 Ke8 25.Bd4 Rc8 26.c3 Qc4 27.Nxg5 Nb4 28.Rxe7 Kxe7 29.Qg7, 1-0.

3) 8...e5 was played in Attwood-Knox, corr. 1982, after 9.dxe6ep Bd6 10.Qh5 b5 11.Bd5 Nf6 12.Bc6 Ke7 13.Nd5 Nxd5 14.Bxa8 Nf6 15.Bh4 Bxe6 16.Bc6 Ne5 17.Bb7 Qb8 18.Bxf6 gxf6 19.Qh6 Qxb7 20.Qg7 Nf7 21.0-0-0 Qb6 22.Rhe1 Qa5 23.f3 Qxa2 24.fxe4 Rd8 25.exf5 Bf4, 0-1. White would have had a superior position after 11.Qxf5.


The speculative 9.Bxb5 has been played in several games. This has led to very complicated positions where Black has done well. Herold-Moriceau, corr. 1990, continued 9...axb5 10.Nxb5 e5 11.dxe6 Bb4 12.c3 Qxd1 (12...Ba5 13.Qxd8 Kxd8 14.Bxc7 Bxc7 15.0-0-0 Ke7 16.Nxc7 Rxa2 17.Nd5 Kf8 18.Kb1 Ra8 19.Ng5 h6 20.Nf7 Bxe6 21.Nxh8 Nxh8 22.c4 Rc8 23.b3 Ng6 24.f4 Bxd5 25.Rxd5 N8e7 26.Rd4 Rb8 27.Kc2 Nc6 28.Rd6 Nb4 29.Kc3 Nxf4 30.Rf1 Ke7 31.c5 Nfd5 32.Kd4 Nc2 33.Ke5 Nde3 34.Rb1 Rc8 35.Re6 Kf7 36.Kd6 Nd4 37.Re7 Kf6 0-1, Munro-Smith,corr 1992) 13.Rxd1 Ba5 14.b4 Ba6 15.a4 Bb6 16.c4 Bxb5 17.axb5 Rc8 18.c5 Ba7 19.Rd7 Bb8 20.Rxg7 (diagram)
N8e7 21.0-0 Nf8 22.Nf4 c6 23.Bh4 Ng8 24.e7 Bxf4 25.exf8Q Kxf8 26.Ra7 cxb5 27.Rd1 h5 28.Rdd7 Nh6 29.Be7 Ke8 30.Bf6 Rg8 31.Re7 Kf8 32.Bg7 Rxg7 33.Rxg7 Rd8 34.g3 e3 35.fxe3 Bxe3 36.Kg2 Rd2 37.Kf3 Bd4, 0-1.

9... Nf6

Black can also consider 9...Bb7 though care has to be taken with the weakened influence on the square e6. For example on 10.Qe2, b4 is now losing after 11.Ba4 Kf7 12.Ng5 Kf6 13.Ncxe4 fxe4 14.Nxe4 Kf7 15.Ng5 Kf6 16.Qe6 Kxg5 17.h4 with mating attack.


After this move Black can play 10...Bb7, which transposes to the game Ottevaere-Vandermeulen above.

Conclusion: These are very rich positions and there has not been enough practise to tell if Black has a satisfactory plan among those presented above. In other words, excellent territory for further investigations.


A note on move order

Black can try to avoid the complications above by playing 6. Bg3 a6 (many of the games above actually had this move order), since Black seems to be doing OK after 7. h4, 7. Nxe4, or 7. f3. Incidentally, most books on the Nimzowitsch defence attaches a ? or ?! to 7.f3, following a Nimzowitsch game where Black played 7...f5. However, if White now plays 8.Nh3 we have transposed into a position discussed above. After 6.Bg3 a6, White’s best is 7.Bc4 after which Black will reach variation B by playing 7...f5 (7...Nf6 is the main alternative). There could, however, be another reason to consider playing 6...a6. In the game Norman-Jensen, corr. 1997-98, White played a new, and what appears to be dangerous, line in 1.e4 Nc6 2.d4 d5 3.Nc3 dxe4 4.d5 Ne5 5.Bf4 Ng6 6.Bg3 f5 7.Bb5 Bd7 8.Nge2 a6 9.Bxd7 Qxd7 10.h4 (this appears to be a new move) h5 (questionable;10...e5 looks more logical but 11.dxe6ep Qxd1 12.Rxd1 Bb4 or 11...Qxe6 12.Nd4 Qd7 does not look satisfactory)11.Nd4 f4 12.Bh2 Nxh4 13.Ne6 Nf6 14.Kf1 Ng6 15.Nxe4 Rc8 16.c4 Nxe4 17.Qc2 Rh6 18.Qxe4 Kf7 19.Nc5 Qd6 20.Nxb7, 1-0.