Black To Play And Draw

IDLY going through old games this morning while waiting for a flight to Amsterdam (to play football, not chess), I came across this interesting minor-piece ending.

Black to make his 55th move in Spanton (1977) - Christopher Archer-Lock (2053), Paington 2012

*****
*****
*****
*****
*****
*****
*****
*****
The fixed nature of the pawns makes the position easier for White to play despite there being rival pawn-majorities (a factor favouring the bishop). Indeed, the analysis engine Stcokfish10 reckons White is winning, but Komodo9 - correctly, I believe - assesses the position as dead equal.
55...Bb2
This loses, but so does every other move except for 55...Be5! The point of the latter is that it lets Black meet 56.Nf7, heading for the queenside, with 56...Bc7, which draws. Stockfish10 disagrees, but it comes up with a series of irrelevant knight and king moves without making progress.
56.Nf7
The only winning move, but good enough. White will pressurise Black's queenside pawns with his knight and use his king to run the bishop out of moves.
56...Bc3 57.Nd8 Be1 58.Nc6 Bd2 59.Kf2 f4?
This makes the win easy. But after, say, 59...Bc3 60.Ke2 Bb4, White wins with 61.Nxb4, eg 61...cxb4 62.Kd2 Ke5 (62...g5 fails to 63.hxg5+ Kxg5 64.d6 etc) 63.Kd3 Kd6 64.Kd4 Kc7 65.c5 Kd8 66.d6 exd6 67.c6 (67.cxd6 also wins), when Black is in zugzwang.
60.gxf4 Bxf4 61.Nxa5 Ke5 62.Nb7 Kd4 63.d6 1-0