Computer Analysis wrongly marking a move as a blunder

[prateeksaini01]

Exact URL of where the bug happened

https://lichess.org/pSQ5nKXK/black#104

Steps to reproduce the bug

Not sure how to reproduce as this is the first time I have seen such a bug

What did you expect to happen?

Move 52 by black should not be marked as blunder
Move 53 by white should also not be marked as blunder

What happened instead?

Both moves are marked as a blunder with the text for losing/walking into forced checkmate added as annotation

Operating system

GNU/Linux (KDE Neon)

Browser and version (or alternate access method)

Firefox 125.0.3

Additional information

This exact position must've happened many times before on Lichess so I doubt that this bug has something to do with that

[jarro2783]

I have another example, https://lichess.org/OKWzOHps/white#164
It was already mate in 2, the move made by black was marked as a blunder, with a different mate in 2 path suggested as the best move.

[benediktwerner]

@jarro2783 which move are you talking about? The move you linked is not marked as a blunder and none of the blunders seem to match your description.

In general, this happens because games are analyzed in parallel and a different Stockfish worker doesn't always find the checkmate on the previous position. There's not really a simple way to fix this for certain.

[jarro2783]

Oh that's strange, it was black's move 82 when I first looked at the analysis. It seems to have fixed itself.