#6 Gligoric-Kotov (click here to view on chessgames.com)
Early in the game, Bronstein finds fault in Black's opening strategy. White has chosen a
somewhat tame variation of the Sicilian, and here Black may have had conflicting desires both to put a bishop on e6 to force (on capturing a White piece on d5) White to recapture eventually with his e4 pawn, blocking the d6 weakness; and also to expand on the queenside as in sharper opposite-side castling variations. Normally this is a reasonable approach, however by the fianchetto of his king's bishop White has already conceded to capturing on d5 with his pawn, as otherwise his bishop would be better served on c4.
Bronstein: The b7-b5-b4 push makes sense only when Black has a pawn on e6 and the knight is left without a good retreat. But here, ...b5-b4 will sent the knight to the excellent square d5, where Black will be practically forced to trade it off, extending the diagonal of the fianchettoed white king's bishop.
9. O-O Nbd7
10. a4 b4
11. Nd5 Nxd5
12. exd5 Bg4
13. Bd2 a5 (Bronstein suggested 13. a5 to prevent ...a5, fixing the b4 pawn as a target before playing Bd2)
14. c3 bxc3
15. Bxc3 Qb6
Bronstein: White has an extra queenside pawn, but making a passed pawn with it would appear to be the last thing on his mind; he wants to do only what is necessary on the queenside in order to free himself to proceed with his attack on the kingside.
16. h3 Bh5
17. Kh2 Be7
Bronstein: Gligoric should have converted his accumulated advantages into a direct attack, beginning with 18. g4, kicking back the unfortunate bishop before anything else; after 18....Bg6 19. f4 would have been considerably stronger.....this would lead to a sharp game, admittedly with some risk to White....but certainly he would also have been left with most o
f the chances.
[Bronstein suggests 18. g4 Bg6 19. f4 f6 20. f5 Bf7 21. Ng3 and perhaps O-O 22. Qf3 with 23. h4 to follow, see analysis diagram - This would have been a very committal decision and even still my computer slightly favors Black, though the momentum is clearly on White's side. It's not hard to understand why White might prefer a quieter way, though in making simple moves he loses the initiative].
Analysis diagram after 22. Qf3
18. f4 Bxe2
19. Qxe2 Bf6
Here White's best plan might be to just open the center with fxe despite his slightly exposed king, since Black's king would not be too comfortable either with a potential exchange sac on f6. Instead, White meanders a bit, as if waiting for Black to fall apart.
20. Qc4 O-O
21. Qc6 Rfd8
22. Rae1 Qb8
Bronstein: ...as soon as Black gets the chance to play ...exf, forcing the trade of the dark-squared bishops, all White's weaknesses will be laid bare.
23. Rb1 Ra7
24. Qc4 Rc8
25. Qe4 Qb3
26. fxe5 Bxe5
27. Qf5 Rf8
28. Qf2 R7a8
29. Qf5 Qxa4
Now, down a pawn and with no compensation, White tried to change the course of the game by an exchange sacrifice and lost in a further 12 moves (See link for the full game). White had built a strong position earlier and was faced with equally tempting choices of trying to attack on the kingside or creating a passed pawn on the queenside. In the end, he chose neither and accomplished nothing - 7 of the last 10 moves shown by White were with his queen, capturing nothing, taking the longest route from e2 to f5. Sometimes the hardest decision to make when playing with an advantage is to make a decision.