Puzzle #2207- Just plain wrong

Puzzle #2207, says the correct move is to threaten white queen with black knight, then queen takes the other knight. It's definitely not forced nor even recommended the queen take the other knight, which allows a forked attack to nab the queen. I set up this position against stockfish 8 and it does not take the knight but retreats the queen to E1 thereby avoiding the fork altogether.

Yes, but then you keep the knight on d4. The puzzle is just basically trying to ask you to try to prove that you can keep both knights on g4 and d4 alive.

This is one of the cases where the move in the puzzle is not the best according to the computer but is the best move to test that you understand the point of the puzzle. The puzzle should be left how it is.

EDIT: oops, i didn't see the position correctly; what happens when the queen retreats is that you just take the free bishop on c4. but still, the puzzle move is better because it is a more strict test to make sure that you have actually solved the puzzle

The puzzle isn't wrong. The whole idea of the puzzle is that the queen cannot take on d4 because of the discovered check. It is far more difficult to find Nxf3+ than Nxc4, which is why the computer should test you on that.

I agree with you there, It does leave the bishop hung. And stockfish also chooses to move the knight there as well. So I can't argue that the first move is the best move. When first viewing the puzzle I saw that as well but to me it seemed that it would be very complicated to successfully defend the knight after that. Playing it out in stockfish yielded quite a bit of complicated play after with a lot of pieces moving around and no immediate exchanges. After considering whites opportunities for counter play, I had cast aside that line thinking the puzzle was trying to show something much more decisive. But I see your point and appreciate your feedback. And after seeing it from that perspective, i'm going to change my post from "just plain wrong" to just a solid down vote. I still don't like the puzzle. :)

And I realize now I can't find how to change or edit my post, so I'll keep working on that. My first response was to "qkxwsm".

Lego00, i'd have to say I disagree with you on that one. I don't think the whole point of the puzzle was to move the knight hoping your opponent would then make any number of bad moves that would drop a queen. I think the point was to kick the queen and gain a bishop without any immediate consequences. The puzzle showing the bad queen move misrepresents whites true potential for counter play in that position in my humble opinion.

Of course, the point is not to "hope" that your opponent takes on d4 so you can win the queen, but that Ne5 simply wins a piece, because white is unable to take the knight without losing the queen. I guess a different way to say it is that Ne5 would be really obvious, if the knight on d4 wouldn't be left undefended. The main difficulty of the puzzle is the discovered attack.

The difficulty for me was not in seeing the queen/king fork. It was trying to calculate far enough down the line after the queen retreated to know whether the bishop would be free or result in some kind of decisive counter play. If they wanted the puzzle to be about the knight forking the king and queen, I would prefer they didn't give the queen a way out causing you have to ponder the merits of ensuing counter play after taking the bishop.

While it seems that indeed retreating the queen is the better move, those defending the knight capture on the grounds that it makes the problem deeper or more subtle are giving the machine too much credit. The LiChess puzzles are automatically generated, and are composed based on a short and shallow engine evaluation. There are too many puzzles, and too many errors in the puzzle solutions, for me to believe any master-level human is actively interacting. It's just random engine stuff.