Not sure if this is the place to report these
lichess.org/training/65184In this puzzle:
after Nxg5 Nxg4, Stockfish lists playing Nf3 as +3.2 and playing g3 as +2, however playing g3 results in an incorrect answer.
g3(+1.9) is a bad move compared to Nf3(+3.4) since it allows Nxe3
I know it's a worse move, just saying it's still winning. Puzzle shouldn't have 2 possible correct answers.
If you have one pawn on the seventh and one queen against a lone king, you have a winning position. However, even though blundering your queen still results in a winning position, no one would say that it should be a solution to a puzzle. It is always ideal to play the best move, and puzzles help us practice that.
there are a very few puzzles on lichess with more than one correct solution(probably lingering around from the early days of lichess tactics).
it says something like "good move but find a better one, try again", when u didnt played the best one, similiar to chesstempo.
i think the depth of the computer analysis is only 16, when i started looking at the position it started with +1.7 and the treshold of a winning position is similiar to chesstempo 1.75.
so it probably barely didnt meet the criteria of being an alternative solution.
On a deeper analysis, Nf3 actually drops below +3 while the evaluation of g3 continues to grow. I currently have both at +2.7, though with g3 at depth 34 and Nf3 at depth 26.
g3 definitely shouldn't be pubished, then. I went ahead and disabled this puzzle.
After g3 white loses an exchange on f1, after Nf3 white is just a piece up.
But after g3 Nxe3 Qd2 Nxf1 Bxf1, black's king is in trouble: if O-O (Stockfish suggests Kf8, leaving the h8 rook out of play) then Bg2 followed by Bxd5, and white has mating ideas via Qd4-h4-h7, as well as on the back rank if black attacks the bishop.
But Black can play Rd8 and d4, creating a passed c pawn. I think white is much better after g3 but Nf3 is clearly winning.
