In <<Candidates' Matches 1974>> by Botvinnik, Matanovic, Kazic and Yudovich, the move order of the present game is given. This is also the case in <<The Games of Tigran Petrosian>> by Shekhtman.

12) ..., Nd4 black creates some dangerous attacks on the kingside. Note the white e2 pawn is pinned.

14) e3 a good illustration of why Kt-D4 was an illusionary threat. D4 was not a permanent outpost. Now the black Kt is pushed back.

14)..., Cxb4 is forced otherwise black loses a pawn.

20) Qf3, the exchange favors white. As black has no reply to g4 followed by g5 forking the Kt and Bishop. White has an advanced pawn on b5 very useful in any protracted endgame.

22) Qf6! The double threat of Qxe7 and Qh4 is decisive.

What is so strange about this match of 13 games is that only one game is more than 40 moves long, so maybe match stamina had less to do with it, than preparation and good moves during the game.

click for larger view

And now the usually Mt Rushmore-like Portisch starts playing flamboyantly.

A few moves later Petrosian plays 12.b3-b4. He has hardly left his own manor; only one of his pieces has advanced beyond the second rank.

click for larger view

But this position is already difficult for black.

Then white plays a few pawn moves, pushing back the black pieces. After 16...Ne7 white's pieces are in the same place they were at move 12...

click for larger view

Petrosian shows his skill here.

Kramnik:

Kramnik vs B Adhiban, 2016

And Carlsen - Gupta referred to earlier:

What was the point of 11 ... Bh6?

My guess is that Black had intended 10...f5 but noticed an apparent problem with it after White played 10. Nd2.

After 10...f5, the natural reaction 11. f4 might have led to 11...exf4 12. gxf4 Bd4+ 13. Kh1 Qh4 14. Qe1 Qxe1 15. Raxe1,

click for larger view

when Black could choose the solid 15...Bd7 or the adventurous 15...Be3. I haven't found a path to a White advantage after either move.

However, instead of 11. f4 (after 10...f5), White could continue 11. Nd5, intending 11...f4 (To prevent 12. f4, which might have been good, now that Black's bishop could no longer reach e3 via d4.) 12. e3 (Not strictly necessary, but it does prevent 12...Bg4, 12...g5, and 12...Nd4.) 12...Be6 13. Ne4 (With 14. Bf3 looming.),

click for larger view

when Black perhaps judged that his kingside attacking chances wouldn't sufficiently compensate for his weakened light squares in the center.

Perhaps this is why Black refrained from 10...f5.

<<N.O.F. NAJDORF>: What was the point of 11 ... Bh6?>

In conjunction with 12...Nd4, it seems an attempt to create tactical threats such as 13...Bxd2 14. Qxd2 Nb3.

<<N.O.F. NAJDORF>: Why did black not play 12 ... Rb8, which would at least have made use of the bishop on h6?>

On 12...Rb8,

click for larger view

13. Qa4 wouldn't yet be playable (which I surmise was your observation). However, Black might have been concerned about the weakness of his c-pawn. For example, 13. bxc5 dxc5 14. Nb3 Qe7 15. Ne4 b6 16. Nexc5 bxc5 17. Bxc6 Rb6 18. Bb5,

click for larger view

and White b5-bishop is not ideally placed. But I'm not sure how Black can exploit this, and he is a pawn down and has three pawn islands to White's two.