1.d4 Nf6 2.c4 e6 3.Nf3 d5 4.g3 The usual Catalan, one of Kramnik's speciailities, with white pieces!
4...dxc4 5.Bg2 Nc6 6.Qa4 Bb4+ 7.Bd2 Nd5 8.Bxb4 Ndxb4 9.0-0 Rb8 10.Na3 Clearly the most fashionable move.
10...0-0 [The tempting 10...a6 seems to lead by force to the following continuation: 11.Ne5 0-0 12.Nxc6 Nxc6 13.Qxc4 Qxd4 14.Bxc6 Qxc4 15.Nxc4 bxc6 with a slightly better endgame for White according to practise, due to Black's weak pawn structure and rather bad pieces coordination.]
11.Qb5 b6 12.Qxc4 Ba6 13.Nb5 Qd5 14.Qxd5 Nxd5 [14...exd5?! is just wrong in my opinion, as it is positionally weak and there is no tactical reason to play it: 15.Nc3 Rfe8 16.Rfd1 Ne7 17.a3 Nbc6 18.b4 Bb7 19.Rdc1 and White was better in Huzman-Cornette, Montreal 2008.]
15.a4 Na5 16.Ne5 Rbd8 17.Nxa7 [17.Bxd5 has already been tried, but it looks like Black simply equalized by force: 17...Bxb5! 18.Bxe6 (18.axb5 is more or less the same sketch: 18...Rxd5 19.b4 Nb3 20.Nc6 Nxa1 21.Rxa1 Ra8 22.Ne7+ Kf8 23.Nxd5 exd5 24.Rc1 Rc8 25.Ra1 Ra8 26.Rc1 Rc8 27.Ra1 1/2-1/2 Rodshtein-Macieja, Copenhagen 2010) 18...Be8 19.b4 fxe6 20.bxa5 Rxd4 and after more and more simplification, draw was agreed in Grischuk-Gelfand (Sochi 2008) later on.]
17...Nb4 18.Rac1 Rxd4
19.Rxc7N There comes the novetly! Actually, the impression I had is that both players had more or less analyzed the rest of the game. After Kramnik's perfect defense, Anand took some time in order to try to find out some ideas, but could not disrupt the solid leader. [19.Nb5 has been played twice, and twice gave White an edge. But according to Rybka 4, Black can defend with 19...Bxb5 20.axb5 f6 21.e3 Rdd8!N (21...Rd2 22.Nf3 was better for White in Prohaszka-Csonka 2010.) 22.Nf3 Rf7 and Black seems to hold in spite of his slightly misplaced knights on the border.]
19...Bxe2 20.Rfc1 f6! Looks like Kramnik knew that the following simplifications are leading to a draw.
21.Nec6 Naxc6 22.Nxc6 Nxc6 23.R7xc6 Rfd8 24.h3 R8d6 This is where Anand spent the most of time (about 25 minutes). Taking into account that Kramnik had more time on clock at this moment, the most logical explanation would be that Anand couldn't find any "real try" at home. After having a new look at the position over the board, he probably drew the same conclusion.
25.Rxd6 Rxd6 26.Rc6 [26.b4 was the other logical try, and I guess Black should hold after 26...Rd4!? 27.Rc2 (27.Re1 Rd2 28.Ra1 Kf7 29.a5 bxa5 30.bxa5 Ba6= ; 27.Rb1 e5= ) 27...Bd1 28.Rc8+ Kf7 29.a5 bxa5 30.bxa5 Ra4 31.Ra8 Ra1 32.Kh2 Be2 and I believe White has no winning chances, as it is very difficult to bring the so-needed King help the a-pawn.]
26...Rxc6 27.Bxc6 e5! Maybe not the only move to hold, but an important one from the technical aspect. Black should gain some space and at the same time not let White go f4 and create Black and isolated pawn on e5 in the event of ...e5 move.
28.f4 exf4 29.gxf4 Kf7 30.Kf2 Bc4 31.b4 g5 32.fxg5 fxg5 33.h4 Setting a definite draw.
33...gxh4 34.a5 1/2-1/2