More endgame precision

Was looking around some files and stumbled upon an Anderssen-Kieseritsky rook endgame from London 1851. I then found something that surprised me. When I finally understood, I thought it might be worth sharing. The following position is from analysis following White's 53rd move. Blacked erred in playing 53 … Kf4. Had he played 53 … Kxh4 the following position would have eventually been reached:

Anderssen-Kieseritsky

London, 1851 (m1.2)

What surprised me when I saw this position is that Black has one and only one winning move. I thought it was either a draw to begin with, or that any reasonable move by the Black rook on the c-file would win. Not so.

I will put the rest below the fold for those that want to work it out on their own.

The only winning move is 1 …  Rc8. The following line makes it clear why this is so: 2 Kd5 Kf3, 3 Kd4 Ke2, and the White king is shut out from attacking the c2 pawn. But if the Black rook were on c4 or c3, the White king would gain a tempo on his march to d3, winning the pawn for free. And 2 Kd7 Rc4 (or Rc3) will keep the White king too far away from the c-pawn to win it. Thus the White rook must eventually give itself up for the Black c-pawn, and the White king will be too far away from the a-file to support his a-pawn's attempt to queen.

These tricks are easy to see with time on the clock and a willingness to calculate, but can be missed easily due to time pressure or thoughtlessness.