Puzzles with inferior solutions

lichess.org/training/62030
has a mating opportunity that's superior to the accepted answer.

After taking the knight with the bishop, the computer responds by taking the bishop with his second knight.

This leaves mate open by plunging the queen to d8, forcing the rook to take, which can then be taken by your rook. Mate.

While the accepted solution gains material, nothing is better than winning the game in terms of solving puzzles.

What can be done to correct this puzzle's answer? Am I missing anything?

Thanks,
Brent

24... Qe8.

I have played over 8000 puzzles and can certainly attest to seeing some that do not show the strongest move,
but we have to remember these puzzles are all taken from actual games on Lichess, so blunders & inaccuracies will appear, to complete the puzzle you must match the move played in the game, which is not always the strongest move.

Press the space bar to turn on the engine to see if you missed anything.

@synposis
That's not correct. You must play the move that Stockfish thought was strongest (by a significant margin) when it analyzed the position, and the opponent will be playing Stockfish moves as well (the moves will sometimes look nonsensical because Stockfish will prefer to give up a rook in one move over losing a queen in a complicated tactic, or sacrifice anything and everything to delay mate). I'd say that the majority of tactics above 2100 were missed or at least misplayed in the games.

The analysis isn't particularly deep, so it sometimes gets things wrong (missing a second winning move; overestimating an attack; not looking deep enough to see a perpetual check; misevaluating an endgame that it would normally handle by much deeper calculation). But this is very rare.

@BrentShaub It's obviously not mate after 24. Qd8+ As @Hastromil pointed out 24. ... Qe8 stops the mate and white doesn't win material as he would after 24. Nb6.

Great. Thank you all for your input.

I failed to see the guarded block, so thanks for pointing that out.

Regarding what determines what the answer is, if it's determined there's a better move than the solution, I wish there was some way to 'flag' what option appears better, so that the puzzle's answer can be changed upon review.

I get so much out of playing these puzzles, and want them to be as skill-building as possible.

Many thanks for the prompt responses. Great community.

Pretty sure the engine Stockfish doesn't make mistakes that you would see :) You have to understand that this puzzles were generated by an engine from games analyzed by Stockfish. Stockfish would never choose gain material over checkmate.

@BrentShaub

Here's the solution: In circumstances where you disagree with a computer engine about what the best move is, assume it is you who is wrong, and go about trying to find out why, rather than going on a futile quest to prove the computer wrong.

There are circumstances where certain people might be able to identify things the computer overlooks. You're not one of those people, and the circumstances that would allow it anyway are insanely rare and unlikely to come up in these puzzles.

The point of the puzzles is to learn and improve your chess, not to prove that you are better than the computer which generates them. Because you're not.

@BrentShaub
You can click the link to the real game after attempting a puzzle that you think is wrong and try your move; Stockfish will inevitably refute it, and by seeing the refutation for your attempt, you'll learn from it as well.

@doom12384 Again thanks for pointing out that it's possible to switch into learn mode when doing puzzles. Spacebar as suggested above was not something I was aware of.

Yes, it makes for great learning to see what Stockfish's ideas are. I've been learning from that already.

@Chessty_McBiggins I don't appreciate your comment about the computer always being better than me. Ya never know.