Continuing on the top 20 of Zurich....
Here it is Black to move, and if I had the White pieces here I might be impatiently waiting for my opponent to resign. White has a ready-made passed pawn and Black's foursome of isolated doubled pawns do not make a good impression.
Bronstein: ....A further demonstration of White's possibilities is that, were it his move here, the 32. c4, followed by 33. Rc3, would leave Black in a completely hopeless predicament, with the rook on c3 simultaneously attacking the pawn on e3 and supporting its own passed pawn. Black's next move is directed against this threat.
Who would guess that White might be barely clinging on to a draw 12 moves later? Credit Gligoric's resourcefulness....
31. .... Re4
32. c3 Kg7
33. Kc2 R4e5
34. g4? h5
Amazingly, a key mistake has already been made, and Black sees a path out of the woods.
35. Rd6 hxg4
36. hxg4 Kh6!
Bronstein: ...but the black king refuses to be tied down! Black is giving up both his f-pawns in order to reach the square f3, which will secure him the draw. Might this possibility have been prevented? Yes; the mistake was 34. g4, by which White transferred the base of his pawn chain from g2 to g4 - i.e., closer to Black's king.
Bronstein goes on to point out a possible winning line for White after 34. g3! which I would summarize as a general concept - White should willingly part with his own f-pawn because if Black takes it then White can take the d3-pawn after which he can start pushing his passed c-pawn with the support of his king and rooks while Black's king is cut off on the kingside. I can empathize with Boleslavsky's mistake of 34. g4 - why not simply protect the pawn, and go from a winning position to a 'really winning' position? This game should remind us to defend only those pawns that are worth defending in an endgame - the pawn count may not matter nearly so much as the activity of the pieces (though ironically White loses all his kingside pawns anyhow).
37. Rxf6+ Kg5
38. Rxf7 Kxg4
39. Rd7 Kf3 (not 39. Ra7?? when 39.... Kf3 leaves White no way to stop the passed pawn)
Bronstein: Black's king has achieved a stunning success, becoming the first to break into the enemy camp.
40. Kd3 Rxf5
41. c4 a4
Here the game was drawn after adjournment - as much as Black has accomplished he cannot win, as White can 'stand pat' and mark time by pushing his free rook along the d-file. The key for White is to resist moving the c-pawn (wouldn't this be irresistible in G/45, let alone a faster game?).
Bronstein: White must move his rook back and forth along the d-file. After 42. c5 he would lose, and most instructively. White's king must defend the rook on e2 and he pawn on c4 at the same time; the latter pawn protects the supremely important support point at d5 for the rook, thus defending the king against checks on the d-file. 42. c5 would allow Black to double his rooks on the fifth rank and drive off the white king, this: 42. ... Ree5 43. c6 Rd5+ 44. Kc4 Rc5+ 45. Kd3 Rfd5+ 46. Rxd5 Rxd5+ 47. Kc4 Rd2 and Black wins....as long as the pawn stands on c4, the king on d3, and the rook covers the d-file, White can hold the draw.
The interesting part to me was the willingness with which Black sacrificed his weak f-pawns to activate his king and secure the draw - I think this idea applies in many contexts.