A beautiful destruction and a rare concept in the Caro Kann Two Knights

Once installed on an outpost in the center a knight can become an incredibly powerful piece to support our attack and interfere with the enemy’s forces.

Due to their long range impact the bishops usually stay behind. They can be equally powerful from a distance.

This game is differenct. For a long time a bishop installed in the center is the most powerful, crucial piece on the board. It tightens the black defense and decisively supports the counterattack. White would have loved to drive it away, but that turned out to be not that easy.

The annotated game below is part of the BodenseeBase for you to download and replay more conveniently.

Villain – Hero

Correspondence game, Lechenicher Schachserver, 2016

1. e4 c6 2. Nf3 d5 3. Nc3


The Caro Kann Two Knights.

3… Bg4

The classical move.

(Lately 3…dxe4 and 3…Nf6 have been played a lot.)

4. h3 Bxf3

(4… Bh5 leads to deeply analyzed, forced variations in which Black may make a draw, but can never hope for more.)

5. Qxf3


The starting position of the Two Knights variation with 3…Bg4. White gets his pieces out quickly and remains flexible with the d pawn. He enjoys the two bishops and some prospects of a kingside attack which can quickly become dangerous if Black doesn’t react carefully. White scores well with this in tournament practice.

When fighting for the World Championship in 1960 the great Mikhail Tal tried 5.gxf3?!! once in order to confuse Mikhail Botvinik. Creating imbalances this way may have been a decent try (once, at least) against the pragmatic Russian chess professor, but the move itself is not so good.

5… Nf6 6. d3 e6 7. Bd2 Nbd7 8. g4


Immediately rushes towards the black king, asking Black: „Where are you going to castle?“ But where to put their kings is the major issue for both sides in this line. Black’s main tries against 8.g4 are either stopping g5 via 8…h6 (and even advancing on the kingside himself) or 8…Bb4.

8… b5!?

Answering White’s kingside attack with an immediate counterattack on the queenside. From now on both sides march on different wings with their kings still in the center. With 8…b5 Black enters unexplored territory. White scored 7:1 in OTB play against 8…b5 according to my database, but it had once been tried by strong British GM Matthew Sadler, so I took a closer look. While I couldn’t find anything wrong in the opening I loved the fact that Stockfish didn’t seem to understand the resulting positions, probably because of their castling and king safety implications.

(8… h6 9. Qg3 d4 10. Ne2 Qb6 11. Rb1 g5 12. Bg2 Be7 13. O-O h5 is how two 2.700+ GMs handled this in Morozevich, Alexander – Vitiugov, Nikita 1/2-1/2 (42 moves) RUS-ch 67th, 2014)

9. g5 b4

The race begins.

10. Nd1 Ng8 11. Ne3 Bd6 12. h4 Ne7


13. h5

(13. Ng4 looks natural, has been played by 2.650 GM Ivan Saric against a lesser opponent. Black should answer 13…Ne5 in order to secure the crucial a1-h8 diagonal for his bishop, the crucial concept also in our game as well as in the Sadler game that had inspired me to try this.)

13… Rb8


14. h6


(14. Qg2 Be5 The Be5 is the central piece around which Black’s play evolves. It protects the dark squares on Black’s kingside, pressures the White queenside (prevents White from 0-0-0 even) and hinders d3-d4 and f2-f4 advances. 15. Rb1 (Had White gone 15.0-0-0? his king would soon find himself in the eye of an irresistible storm.) 15…Qc7 16. Be2 a5 17. Rh4 Rb6 18. Qh3 Bc3!? with an unclear game in Hall-Sadler, 1-0 (60 moves), Bundesliga 2002.)

14… Be5


It’s a rare concept that a bishop instead of a knight becomes installed on a central square on a full board. In this special case (like in the Sadler game) the dark squared bishop is the one piece that ties the black army together. It defends the dark squares on the kingside and pressures White’s queenside at the same time. In fact b2 is hanging right now. As long as Black can keep the bishop on its exposed outpost he will be fine. And it it is hard for White to chase it away.

15. O-O-O

Implies a draw offer already since Black can now force a perpetual check! Otherwise a dangerous choice. I doubt the White king will ever be safe after castling into the already running black queenside attack.

(15. c3 is what I expected, fighting the dominant e5 bishop. 15…bxc3 16. bxc3 d4 17. cxd4 Bxd4 with unclear consequences.)

15… Qa5


16. Kb1

The only move.

(16. hxg7? invites a powerful rook sacrifice. After 16…Qxa2 17. gxh8=Q+ Bxh8 18. Be1 Bxb2+ 19. Kd2 Ne5 Black has more than enough compensation for the rook.)

(16.a3? not only looks dubious, it runs into the immediate 16…Lxb2+! and White will be mated.)


16… g6

Closing the kingside, refusing the draw offer. For a human it seems logical and natural to close things down where White is attacking, but the engine disapproves. Stockfish prefers 16…gxh6, a move I didn’t seriously consider. After 16…g6 there are some dark squared wholes in Black’s camp, though, emphasizing the importance of the black squared bishop that needs to cover them for now.


(16… Bxb2 forces a draw. 17. Kxb2 Qa3+ 18. Kb1 b3 19. axb3 Rxb3+ 20. cxb3 Qxb3+ 21. Kc1 Qa3+ 22. Kb1 = (22. Kc2 If White tries to escape the perpetual check he gets mated in the middle of the board, and beautifully so: 22…Qa2+ 23. Kc3 (23. Kc1 Nc5 24. Re1 Nb3+ 25. Kd1 Qxd2#) 23… d4+ 24. Kxd4 Qb2+ 25. Kc4 Ne5+ 26. Kc5 Qb6+ 27. Kd6 c5+ 28. Kxe5 Ng6#, diagram))

(16… Rb6 Here and almost during the whole game …Rb8-b6-a6 is desirable, bringing another piece to the attack. Rook up and over? No, White can always parry the rook transfer with d3-d4 so that the f1 bishop controls a6. 17. d4! Bxd4 18. Bc1 and White is better. He can now tear things up around Black’s king while the black attack has come to a standstill. The Rb6 looks silly suddenly.)

17. Ng4


17… Bc3

The bishop needs to stay on the long diagonal, doing his multipurpose offensive and defensive job. c3 is the perfect square from where it cannot be driven away.

(17… Bd4 Was a good, tempting alternative. On d4 the bishop is less stable than on c3, but it takes away the d3-d4 option from White and powerfully threatens …Rb8-b6-a6.)


18. Nf6+?

This feels very wrong, closing Black’s weak point f6.

(18. Bc1 White should buy himself some time, consolidate the queenside with 18.Bc1 and then look for something active.)

18… Nxf6 19. gxf6 Nc8


Black is in the driver’s seat now, White lacks active options.

20. Bc1?!

Too late for consolidation. White desperately needs counterplay, but now Black gets all his forces rolling towards the White king, and White has no measures of relief anywhere.

(20.exd5 was a better attempt to stirr things up. Black has the choice of playing it calmly with 20…cxd5 or enter messy complications after 20…Bxd2 21. dxe6)

20… Nd6 21. Qg3 Kd7


22. d4?!

Hard to criticize this since it was difficult to find a move.

(22. Bf4 may have been the last chance. 22…Nb5 23. d4 (23.Bxb8 Bxb2 and Black wins) 23… Na3+ 24. Kc1 Rb7 Black is much better, having the white monarch under fire, but there is nothing immediate and things remain somewhat unclear. White may have a final opportunity to look out for ideas on his own at this point.)

22… Nxe4

23. Qe3



looks spectacular, threatens …Nc3+ winning, but in the end this is just pawngrabbing on f2 after White has defended against …Nc3+.

(23… Bxb2 Stockfish likes this, but it seemed less clear to me. 24. Bxb2 Nc3+ 25. Kc1 Nxd1 26. Kxd1 Qxa2 27. Qb3 Qxb3 28. cxb3 A funny configuration: Black’s rooks are handicapped by their own forces. Because Black still has all eight pawns on the board the rooks have trouble finding open files. White might be able to set up some fortress and control all intrusion squares forever.)

24. Qb3

(24. Rxe1 Nc3+ 25. bxc3 bxc3+ 26. Ka1 Qb4 -+)

24… Bxf2 25. Ka1 Qc7 26. Rh3 Qb6 27. Qd3 Bg1 28. Be3


28… Bxe3

Goodbye and thank you, proud bishop! You did the most amazing job in this game.

29. Qxe3 a5 30. Bg2 Nd6 31. Qc1 Rhc8 32. Rg3 a4 33. Rg5


Pretending there might be Bxd5 business if Black goes 33…c5.

33… c5 34. dxc5 Rxc5 35. Bf1 b3 36. Rg2 a3 


A picture of utter and complete black triumph. Now White missed the most aesthetically pleasing spot to resign the game and made one more move (37.c3) instead. But since the winner gets to write the history I’d like to pretend that this is the final position of the game.


0 0 vote
Article Rating
Benachrichtige mich bei
Most Voted
Newest Oldest
Inline Feedbacks
View all comments

[…] me 3…Bg4 works well, and just lately it has awarded me with a beautiful win in a correspondence game you already know. Just for the sake of fun and variety I may give […]


[…] 2015 am Brett ausgeklügelt hatte. Nun, MVL-approved, hat Schneiders Zug die Theorie der Zweispringer-Variante […]