I have noticed that if occasionaly I take back a move that maybe an innacuracy or mistake and replay the same move it changes to good or excellent. And if I do it again it changes back to an innacuracy. If I make a move that is good and replay the move it may change from best to excellent or excellent to best I am playing level 7 (1000) AI bots. I think this maybe a bug except sometimes I think the move is the best move? It is easily repeatable.
Hi @cmu3rd it is expected that after a take back that the feedback would change. Stockfish does have variance each time it runs so it will produce different results for the same position. If you have any particular position that the feedback seems wrong please provide the position and a picture of what the feedback shows. Then we can take a look at it. Thanks!
If you find one that went from mistake to excellent - that is one we would like to see. The stockfish randomization can be expected to shift things slightly (i.e. mistake to inaccuracy), but if you come across a case where a move drastically changes, we would like to confirm what happened. Thanks