1.Nf3 d5 2.d4 Nf6 3.c4 c6 4.Nc3 a6 The Chebanenko Variation. Frankly speaking, I cannot understand why he classical players of the early 20th century could not find this useful move, keeping the possibility of actively developing the traditional problem child bishop on c8. Nowadays, the move seems so logical and natural!
5.a4 One of many sensible replies. What hasn't White tried, in an effort to refute the Chebanenko! Yet even the greatest sceptics have to acknowledge that Black's play has a healthy positional basis. The text move prevents Black gaining space on the queenside by b5.
5...e6 A solid response. True, the bishop on c8 remains passive after all, but White has had to pay for this achievement by weakening the square b4. [The drawbacks of 5...Bf5 6.Qb3 Ra7 7.a5 were revealed in the game Gelfand-Morozevich, Astana 200: 7...e6 (7...dxc4 is more reliable) 8.Qb6! Qxb6 9.axb6 Ra8 10.c5 Nbd7 11.e3 Be7 12.Be2 0-0 13.Nd2 e5 14.0-0 Rfe8 15.Nb3 Bf8 16.Bd2 Bc2 17.Na5 and White's pressure on the queenside was overwhelming.]
6.g3 Magnus heads for a Catalan set-up. [The latest tests show that after 6.Bg5 Nbd7 7.a5 (7.e3 Qa5! ) 7...h6 8.Bh4 dxc4 9.e3 b5! 10.axb6 Nxb6 11.Ne5 c5 Black successfully equalizes.]
6...a5 Securing a favourable queenside pawn structure. The black pieces will feel comfortable on the b4 square.
7.Bg2 Nbd7 [A rare plan. Usually, Black directs the knight to b4 via a6, eg. 7...Na6 8.0-0 Be7 9.b3 0-0 10.Bb2 b6 11.Ne5 Nb4 12.e4 Ba6 13.Re1 Rc8 14.Rc1 dxc4 15.Nxc4 Bxc4 16.bxc4 e5! and Black took over the dark squares in the centre (Ruck-Rustemov, Warsaw 2005).]
8.0-0 [A novelty. In most cases, once Nbd7 has bene played, White has to defend the c4 pawn, but Magnus is ready to sacrifice it. 8.Nd2 Be7 9.e4 dxe4 10.Ndxe4 0-0 11.0-0 Nxe4 12.Nxe4 Nf6 13.Be3 Qc7 14.Qb3 Ng4 15.Bf4 e5 16.dxe5 Nxe5 17.Qc3 f6 18.c5 and White was better (Pham Hoang - Le Quang, Vietnam 2003). However, Black's play can easily be strengthened, eg. it would be interseting to try 9...c5!]
8...Bb4 [So this is why the b4 square was left free. 8...dxc4 9.e4 Bb4 10.Qe2 Nb6 11.Bf4 Black will have to suffer for a long time in a passive position.]
9.e4 [A sharp reply! Carlsen opens the centre at once, exploiting Black's slight lag in development. 9.Nd2 ]
9...Bxc3 [The most principled. less convincing is 9...Nxe4 10.Nxe4 dxe4 11.Ng5 and now 11...f5? is impossible because of 12.Nxe6 ; Also 9...0-0 allows White pressure after 10.e5 ]
10.bxc3 Nxe4 [Worse is 10...dxe4 11.Ng5 and White regains the pawn at once.]
11.Ba3! Having said "a", White must say "b". He needs to create threats, not worrying about the c3 pawn. Carlsen cuts off the enemy king in the centre.
11...Ndf6 [Kamsky does not grab the second pawn. After 11...Nxc3 12.Qb3 Ne4 13.Rfe1 White certainly has very strong pressure. One cooperative variation runs 13...b6? 14.cxd5 cxd5 15.Nd2 Nxd2 16.Qxd5 Ra7 17.Rxe6+! mating.]
12.Ne5 White must proceed energetically.
12...Nd6 Kamsky's idea is simple - he closes the a3-f8 diagonal, and so prepares castling.
13.cxd5 exd5 14.c4 Still trying to open central lines at all cost.
14...0-0 Kamsky does not even look at taking the pawn. The safety of the black king is worth all the pawns in the world.
15.cxd5 Nxd5 White has compensation for the pawn.