25.Nh3 is Houdini 1.5's first choice. It points out Black's best defensive set-up to the threat of Qh6 and Ng5: ...Kh8, ...Qb8/d8 and ...Qg8.

Why not make that move?

I think Anand intended 16...Nxd5 in reply but I am not sure.

Btw, 15...d5 in itself is probably fine if Black follows with 17...a5! (which, like Rdc8, defends against the threat of Ba5) - then the idea of advancing the e-pawn can be efficiently countered by 18.e5 Nd7 19.e6 d4! and now (a) White doesn't have Ba5; and (b) 20.exd7 isn't played with an attack on Rc8, so it can be met by 21...dxc3.

Gaito: Anand resigned because he knew his position was hopeless. Notice that in the final position almost all of Black's pieces are tied and immobilized, as they have to protect vulnerable squares or pawns. Black has too many weaknesses, especially the dark squares on the K-side. He can hardly find useful moves, and having to defend such an awkward position against Magnus Carlsen might be a kind of psychological torture for Anand or anybody. But spectators would have liked to see the winning procedure. A possible continuation might have been something like this (SF13 vs. SF13 at 30 seconds per move; hardware: 20 CPU core Dual Xeon E5-2680 v2): 30....Qc8 31.Rxd3 a5 32.Kg2 Rb8 33.Rde3 Rbb6 34.Re2 Qg8 35.Qh4 Rxb3 36.Nxe6 Rb7 37.Qg5 Qb8 38.Qh6 Qg8 39.Nxc5 Rbc7 40.Ne6 Rb7 41.c5 Rc8 42.Qd2 (computer evaluation: +7.38). See diagram below:

click for larger view

The position is utterly hopeless for Black.

Gaito: Let us have a look at the position after 23...Ng7 (See diagram below):

click for larger view

Carlsen played a beautiful positional move: 24.g4!, depriving Black's knight from its two most useful squares f5 and h5. Chess students are apt to learn a lot from this move. Great positional masters from the past, like Botvinnik and Capablanca were also fond of playing such pawn moves that restricted the mobility of the enemy knights. But that move is far from being a winning move (computer evaluation is only +0.69). Anand replied with 24...Rc6?, and that was another "little mistake", as the rook on c6 will obstruct his own queen. Sure enough, Anand was woried by the weakness of his e6 pawn, so he added more protection to that pawn. The engines prefer 24...Qd6, so that Black's queen would be closer to the center and the K-side and would not be obstructed by a rook on c6. After 24...Rc6? Carlsen played another beautiful positional move: 25.Nh3! with a view to taking immediate advantage of the weakened dark squares around Black's king. Now that Black's knight has no useful moves left, White plans to move his Q to h6 and his Knight to g5, then place a rook an e5, and Black's position would be completely tied up and devoid of any useful moves. It is a wondrful positional conception in the spirit of Rubinstein or Capablanca: a strangehold of Black's position is being planned, and it doesn't seem that Anand can do much about it.

Gaito: the last critical moment of the game is depicted in the following diagram:

click for larger view

Black can hardly move any of his pieces: touching his knight would be a suicide, and his queen and one rook are used only to fulfill the thankless task of having to protect the weak pawn on e6. Anand played 27...d3?, maybe out of despair, as the alternative was to sit and get slowly strangled. The engine plays 27...Qa6 as the lesser evil, but the evaluation is +2.60. Yet, after 27...d3, 28.Re5! was the last nail in Black's coffin (evaluation: +5.07). This game was really a beautiful positional masterpiece.

Except that in that game there is no way for white to break through. Capablanca vs Alekhine, 1936 is a draw by perpetual static advantage.

Anand maybe gets a little sloppy with ...Rd8 and ...Qb6 and gets a bit thrown by the alert Qe1. Don't think 17...Rdc8 was part of the plan.

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