| Peligroso Patzer: Taking another look at the game's final position after <63. … Ke5> (see diagram in previous post), I am not at all certain that White is losing; he may actually be winning. The tablebase linked in my previous post verifies that if White had continued with the obvious (but likely not best) <64. Rxb3> (as analyzed above), he would have been lost. On further reflection, however, it seems apparent that there is no hurry to take the b-pawn (which is going nowhere with White’s rook on the b-file). Much more to the point is to move the King to the d-file to get within the square of the foremost Black f-pawn. Best seems to be <64. Kd3> (to prevent 64. … Ke4). The resulting position would be (with Black to move): 
click for larger view In the above position (which has seven [7] units, one too many to be in the tablebase, thus leaving us to our own resources), I do not see any way for Black to make progress. If <64. … f4> (to prevent 65. Ke3), the tablebase verifies that White is winning after the simple <65. Rxb3> (although seemingly better in principle would be first 65. Rb5+, to force the Black King to the sixth rank). Pushing the b-pawn (<64. … b2>) seems pointless, and the tablebase verifies that White is winning after the simple <65. Rxb2> (although, again, there seems to be a more principled move, <65. Ke3>, to prevent 65. … Kf4). Pushing the foremost f-pawn is no better. After <64. Kd3 f2>, White must be winning after <65. Ke2> (staying within the square of the pawn). Finally, running the h-pawn (<64. … h5 65. Ke3 h4 66. Kxf3> also yields a position that the tablebase verifies is <  >. In summary, from the final position of the game (as presented in the diagram in my <previous> post, White seemingly would have been winning if he had played <64. Kd3>, which strengthens the hypothesis that White probably lost this game by time forfeit. | 
Kibitzing
sac: 33...h6
>.