Stockfish eval????

In this game, why does the evaluation suddenly drop to ~0 after I played the recommended move on move 57? It seems to be a clear win...

Is it a win though? What if black had gone after your g-pawn and promoted his f-pawn one half-move after you promote your h-pawn? You cannot force a trade of queens it seems, and even winning the black d-pawn might be hard.

I think it is?

I played around with it a little, and Syzygy tells me it's a win if we trade pawns?
Maybe I misunderstood what you meant?

Could you maybe show me exactly which line you meant?

Also I'd like to note that it's somewhat odd for the eval to jump from inf to 0 and back to inf after just playing the recommended moves each turn

On move 57 you have a clear win indeed. He should have moved 56...Ka2 avoiding trade of queens and aiming for repetition of moves to draw. The aberration in Stockfish evaluation is due to the horizon effect. Syzygy is always right, Stockfish often not.

Thanks!

@skovikes It is a win with the correct plan, which is going after Black's d-pawn and promoting your d-pawn. Going after the g-pawn was a bad choice and the only reason you won quickly was because he mistakenly went for your d-pawn.

And the variation I mean is 60... Ke3 61. Kg6? Kf4 62. Kxg7 Kxg4 63. Kxh6 f5 64. Kg6 f4 65. h6 f3 66. h7 f2 67. h8=Q f1=Q when you've given yourself a much harder technical task of winning this endgame, which might just be a tablebase win. This compared to 60... Ke3 61. Ke6! when ultimately you save a move and promote on move 66, while black still cannot promote before move 67, with an easy win.

ah i see @thijscom

thanks for the explanation!

Yeah, it's some sort of SF bug. It sees clear win with Qc8+ & Qxc1, but doesn't see the same thing for Qc6+ & Qxc1