Looking at the post-mortem with the players after the game (video 3.5 in http://www.londonchessclassic.com/, starting from about 20:00 - highly recommended), it's quite clear that Carlsen was really rattled during the game when he realized the mistake he's made with 24...Be4; for some reason he was counting only on 25.Nh6+ gxh6 26.Qd4 f6 27.Re1 Bc5 28.Qxe4 Qxe4 29.Rxe4, with a difficult endgame for White [Anand says that "he didn't see it at all":-)]. So he had to make a very quick mental switch from playing for an advantage and a win to prudently playing for a draw, and apparently he didn't handle it well.

Eyal: Speaking of the post-mortem, here are some more interesting lines mentioned with regard to critical points in the game:

click for larger view

Here Anand mentions 24...Qe6 25.Ncd6 as better for Black, though he wasn’t sure how much – instead of the forcing 25...Rc5, he mentions 25...Rb8, where White cannot regain the pawn - 26.Nxb7 Rxb7 27.Rxa6 loses to 27...Nc5.

click for larger view

Here he gives a nice line to refute 26…Qe6: 27.Nxc8 Nc5 28.Ba3! Nb3 (28...Qxc8 29.Rc1) 29.Qd8 Qxc8 30.Rd1! – and Black can’t do anything to prevent the exchange of queens and the entry of the white rook to the back rank, leaving White an exchange up.

click for larger view

With regard to 28…Rc7, it seems that after the rather obvious sequence 29.Qd5+ Kh8 30.Qf7 the players focused mainly on 30...Qc8 – Anand gives 31.Nh6 (31...gxh6? 32.Rxd7 and 33.Bxf6+) and says he couldn’t find anything clear after a bishop move to c5 or b4, but that Carlsen told him after the game that 32.Qe6 should win, which is indeed confirmed by the engine: 31...Bc5 32.Qe6 and the threat of Nf7+ (and then inflicting some damage by discovered check) is decisive. 31...Bb4 is a bit more interesting, since it allows 32.Qe6 Nc5; White can win here by 33.Qg4! threatening both Qxb4 and Nf7+ followed by Rd8+, and Black can’t exchange queens because of mate on d8. Sample lines are 33...Qf8 34.Qxb4 gxh6 35.Bxf6+! Qxf6 36.Qb8+ and Qxc7, or 33...Rd7 34.Rxd7 Nxd7 (34…Qd7 35.Qxb4 gxh6 36.Bxf6+) 35.Qe6! and again the threat of Nf7+ is decisive (35…Nb6 36.Qxb6, and the knight on h6 is still untouchable because of Bxf6+).

Besides such concrete lines, which can be easily checked with an engine, it’s very interesting to hear the players (well, mainly Anand – Carlsen isn’t very talkative for obvious reasons) evaluating various possible endgames that could have turned up – e.g., Black’s chances of setting up a fortress in case of 31.Rxd7 Qxd7 32.Qxd7 Rxb2.

Another point I like to mention is that I really like Black's position in Breyer (say after 13. ... Bf8). The problem is whenever I meet 1. e4 with 1. ... e5 most of my opponents avoid the main line. They play almost everything else: King's gambit, Scotch, Bishop's opening, 3. Bc4, the exchange Lopez, .... and I don't like to play Black in those sidelines (except 1. e4 e5 2. Nf3 Nc6 3. Bc4 Nf6).

Btw, as far as the opening stage goes, Carlsen has indeed been consistently getting good positions in this Breyer line lately (in all the three games with Anand, as well as in the game with Shirov from Bilbao). The only one who managed to get an advantage against him is Gashimov, with the older 13.Nf1 line.

Well I would not say the "correct lines" but rather "main lines". That was basically the reason I stopped playing 1. ... e5 and started playing 1. ... c5 instead.

http://www.youtube.com/watch?v=_MDW...

/shrugs meekly, humbly, waits for reply... :)

That's an RSS feed, but you can probably find it by looking up Kavalek on huffingtonpost.com

- What he missed while calculating 33.Rf7 was 33...Qc8 34.Qe7 Rxb2 35.Rxg7 Qc1+ 36.Kh2 Qf4+ 37.Rg3! and the rook both stops Black's attack and keeps the deadly mate threats of White.

After 37.Qf7 Rxf2 38.Kxf2 Qc5+ 39.Kf3 Qa3+, Houdini 1.5 suggests <40.Kg4 Qb4+ 41.Kh5 Qb5+ 42.Rd5 Qe2+ 43.g4 Qe7 44.Qxe7 Bxe7 45.Kxh6 Bf8+ 46.Kg5> (+10.25/d=23). What line does your engine suggest after 39...Qa3+?

See something that violates our rules? Blow the whistle and inform a moderator.