Current 4 follow-up questions for the answer (currently just 1 answer):
A - engine says Kf5 is a blunder. Did you mean Kf6?
B - engine says Kf5 Re7 is a blunder. Did you mean Kf5 Rd7? Kf5 Re7 blunders into king and rook mate with Kf6 next?
C - after Kf6 (instead of Kf5), white must not play a rook move right? after Kf6 any rook move leads to rook being captured or an actual king and rook mate right?
D - 'BTW I just autoplayed the position after Ke6, and White doesn't make any progress at all' --> aha! so Ke6 is actually a draw?
Click here for engine re position below
8/p2k4/1p4p1/1P1p2K1/3P1P2/R3P3/8/1r6 b - - 0 1
The way it's described by hanging pawns in A Rook Endgame Lesson with Akiba Rubinstein at 29:09 (the game is discussed from 22:48 until 31:21) makes me think it's some ordinary expected move while alekhine is already losing. but according to engine: there is a good move (or at least a better move), and the only good move here is Ke6. but then the evaluation goes from 0.6 to 3.2 for Rxb5.
the way this game is being described sounds as if alekhine's been doomed from the start of the endgame and didn't have any chances of winning or drawing in between and then it's just a matter of rubinstein to convert the win and it's like rubinstein does this perfectly. but it seems like there could've been some chance for Ke6 instead of Rxb5.
What's going on?
Similar question: Ulf Andersson vs Robert Huebner 1981 rook endgame: the mistake here is Rd8 right?