He had spent 8:44 on the sacrifice 23 Nc5 and answered within seconds up to that move. Then he thought for 7:01, but could not rescue the game.

Life imitates art. I was just reading Nabakov's "Luzhin Defense" and guess who this passage reminded me of:

<Luzhin’s game, which in his early youth had so astounded the experts with its unprecedented boldness and disregard for the basic, as it seemed, rules of chess, now appeared just a little old-fashioned compared with the glittering extremism of Turati.

Luzhin’s present plight was that of a writer or composer who, having assimilated the latest things in art at the beginning of his active career and caused a temporary sensation with the originality of his devices, all at once notices that a change has imperceptibly taken place around him, that others, sprung from goodness knows where, have left him behind in the very devices where he recently led the way, and then he feels himself robbed, sees only ungrateful imitators in the bold artists who have overtaken him, and seldom understands that he himself is to blame, he who has petrified in his art which was once new but has not advanced since then.

Looking back over eighteen and more years of chess Luzhin saw an accumulation of victories at the beginning and then a strange lull, bursts of victories here and there but in general— irritating and hopeless draws thanks to which he imperceptibly earned the reputation of a cautious, impenetrable, prosaic player. And this was strange.

The bolder his imagination, the livelier his invention during his secret work between matches, the more oppressive became his feeling of helplessness when the contest began and the more timidly and circumspectly he played. > (https://archive.org/stream/in.ernet...)

The machine does not care about immediate, direct protection and invited the other rook to the party: 23Rda1. Alternatively, it suggests to play 23Ra8 with the idea of 23...Rb7 24Rda1 Rc7 25Ra1a7! and Black is still in a bind. Cood blooded play for domination.

Well, later I realised it fails as black has 28... Bd5! may be this is what Anand missed when he made the sac?>

I don't understand - how can black play 28...Bd5 in this sequence?

Suppose that instead of 26 Bd5 White plays 26 g3. Then on 26...e4 27 Rd5+ Black's king cannot advance to f4 in order to hold on to the e4 pawn. However Black can make a waiting move after 26...Bf6 27 Ra8 with 27...g6 ( say) and on 28 Rd1-a1 e4 becomes possible as White lacks the check Rd5+

A pawn on the seventh rank seems considerable compensation for a piece if White can make it count. It is however blockaded on a white square.