Now we come to the key variations in the modern game, where the Caro-Kann starts to resemble a Sicilian Defense with opposite side castling and pawn storms. In what Kasparov terms 'Larsen's revolution', it became clear that kingside castling was indeed possible for Black and may, in fact, provide him with more resources.
Starting from our main position after 10. Qxd3:
Kasparov notes that now:
Black most often plays 10....e6 and if 11. f4 - ...Ngf6 and ...Be7, completely rejecting the classical ideology of the variation, which was based on the following dogmas:
1) it is essential to play ...O-O-O, since on the kingside Black is castling into an attack
2) therefore ...Qc7 is essential, since otherwise the white bishop will occupy f4, which with queenside castling is dangerous
3) after completing his development, Black frees his game with ...c6-c5: he simply does not have any other plan.
Today, however, castling on opposite sides has become the norm, since Black has arrived at a new understanding: kingside castling is by no means an act of suicide!
We will see that this is true, yet all the same White has a variety of ways to put pressure on Black's position, ranging from sacrifical attacks to early endgames. The resulting positions are rich with possibilities. In part 2 we will look at non-cla
ssical variations with early ...Qa5+
and/or ....Bb4+ with kingside castling, but in part 1 we will see the main line:
10. .... e6
11. Bf4 Ngf6
12. O-O-O Be7 (12.... Nd5 13. Bd2 Nb4!? 14. Qb3 a5 15. Ne4 a4 16. Qa3! is rated as better for White, 15. Kb1 a4 16. Qe3 was also seen in Leko-Dreev 2002)
And now White has choices depending on stylistic preferences - 13. Kb1, 13. c4 and 13. Ne5.
We'll start with 13. Ne5, which was first thought to be critical and logically follows from Bf4. No discussion should avoid the amusing game Beliavsky-Larsen 1981:
Behold the pale horse. The man who sat on him was Death...and hell followed with him.
This game asks the question - what if Black 'passes' for two moves?
13. Ne5 a5
14. Rhe1 a4? ('Larsen is doing a wonderful impression of a beginner' - Gallagher)
15. Ng6! Nd5 (White even wins after ...fxg6 16. Qxg6+ Kf8 17. Rxe6 Qe8 18. Rde1!)
16. Nf5!! and White won in a further seven moves (see link above, it's fun to work through the variations)
Death rides a pale horse
This game demonstrates a key feature of kingside castling for Black - both sides must always consider the consequences of White dropping knights on f7, g6, or f5. These strategic risks do not exist in queenside castling lines.
However, after the unambitious 13. ... O-O it seems as though White has some trouble maintaining a meaningful initiative against good play from Black
. For example the natural looking 14. Ne4 Nxe4 15. Qxe4 Nxe5 16. Bxe5 Qd5 already gives Black equality and White even lost the well known game DeFirmian-Korchnoi 1989
from this position. Why is this? 13. Ne5 is a dangerous looking move but does not slow Black's development or (against good play) meaningfully advance White's attack as the f7/g6/e7 squares are adequately covered.
13. c4 is another bold method to try to refute kingside castling. This highlights the absence of a Black rook from the d-file amongst other things - d4-d5 is one plan. Also, for once White does not simply allow Black use of the d5 square. But Black, ever inventive, notices a drawback...
13. .... b5!? (I would like my d5 square back please)
14. c5! O-O
now follows Ivanchuck-Seirawan 1990
(see link for Seirawan's own analysis) - Black has the d5 square but White has, importantly, closed off files on the queenside and maintains his space advantage.
15. Kb1 a5
16. Bc1 a4
with complicated play. An interesting follow-up to this game was Ivanchuk-Korchnoi 1993
, where after ...a7-a5-a4, Korchnoi preferred to castle...queenside!
Black can avoid the committing 13 ... b5!? and prefer 13. ... O-O when the lines can transpose to other options described here. However, if not challenged White does best not to commit to piece exchanges and instead try to find ideal placements for his knights and initiate his kingside attack.
The more indirect 13. Kb1
in combination with 14. Ne4 has become popular at the GM level, although this is not to everyone's tastes - the main line leads to a quick endgame:
13. Kb1 O-O
14. Ne4 NxN
15. QxN Nf6
16. Qe2 Qd5!? (combines centralization with an attack on h5)
17. Ne5 Qe4
18. QxQ NxQ
A critical position (!?) In this dry-looking position it turns out there are interesting tactical resources that give White a chance for an edge. For example, after 19. Rhe1! Nf6 (...Nxf2? 20. Rd2 Bh4 21. Ree2! and Black is embarrassed by his wandering knight) 20. g4!? and the endgame was executed to
perfection by White in the stem game Kasparov-Anand 2003
White also has the choice of a different and slightly favorable endgame after 19. Be3!? when the ...c5 break has been momentarily suppressed.
Key source games include Stefansson-S Kasparov 2001
and Kramnik-Bareev 2003
However, if Black is not in the mood to defend a grueling endgame then more interesting play arises from 13. Kb1 O-O 14. Ne4 Qa5!?Again Black targets the h5 pawn, a logical response to removing the knight from g3. However, now White can try a thematic and dangerous pawn sacrifice:
15. g4!? Nxg4
16. Ne5! Ndxe5!
17. dxe5 Rad8
18. Qh3! and this follows Motylev-Willemze 2003 where White converted his attack ultimately into a win, albeit in an endgame.
The fireworks were truly on display in Polgar-Anand 2003, where Anand came up with a shocking novelty in a related line, and Polgar managed to draw only with difficulty.
Conclusion: Kingside castling for Black carries strategic risks, but it is clearly not nearly as risky as was once thought - indeed the main line goes straight to an endgame. However, a patient player from the White side can either take their chances with the slight edge given to them there, or go for more entertaining fare, particularly 13. c4 where the verdict is unclear. It's not easy, but the 'Kann is all about pain tolerance, time to hit the gym and bulk up.