Bishop equals 3 pawns ... are you sure ?

Recently I had an endgame K+B+4P vs. K+6P. Despite of my expectations, the game ended in a draw. I was upset and decided to get to the root of it.  

Bishop equals three pawns, we learned that at the beginning of our chess careers. So, if in the endgame you have bishop for 2 pawns, you got great winning chances, right? Not so simple. Excellent example is the first game from the World Championship match Fisher-Spassky in 1972. After bizarre move 29. … Bxh2

that Fisher did ( some people say, that it was miscalculation, he counted on the following line: 30. g3 h5 31. Ke2 h4 32. Kf3 h3 33. Kg4 Bg1 34. Kxh3 Bxf2 ,

but didn’t see 35. Bd2! catching the bishop) from equal bishop ending it became bishop vs. pawns endgame. He had drawing chances later, one of them was pointed out by Speelman:

Fischer playing 37. ...a6 instead of 37. ...Ke4

37…a6! 38.Kf3 axb5 39.axb5 b6 ! 40.Bf8 g6 41.Bg7 f5 42.Bd4 Kc4 43.Bxb6 g5 ! 44.Bc7 Kxb5 45.Bd8+ g4+ 46.Ke2 Kc4 

with a clear draw. Finally he lost.

As for my game, looks like I didn’t have the right plan in the beginning, then didn’t use the chances that my opponent gave me, the last one:

White played 54. d6? instead of 54. Kb6 with a draw.

Black returned the favor, playing 54. ... Kf7 which leads to draw, instead of playing 54. ... Bg5 55. Kc6 Kf5 56. Kc7 Ke5 57. Kc6 Kd4 with a win.