Stockfish computer's top opening moves actually loses!

Sort:
infinitemeditations
I've been playing around with the stockfish engine on lichess.com, which is cloud computing and very powerful, and found that it's top recommended moves for black is actually losing if white plays the Nimzo Indian! 
This is all done on the engine's top moves for black. Stockfish evaluates the position as 0.0 here. The problem is, when played out, the black Queen gets trapped in all variations and black is losing. 
From here black will eventually have to lose his Queen. 
Strayaningen

It is not even close to true that Black loses his queen in all variations, for starters Black's best move is not 16...Qh3 but 16...Qg6, where White is only a little better. But it is true that 11...Qh3 is (thought to be, in current theory) not a very good move and that you have to run Stockfish to like depth 40 before it sees this. This is a well-known theoretical position where the correct move is 11...Qh5 with equality. There is no cloud entry in the Lichess database there, so all you're doing is running Stockfish to depth 28 only. Also it is Stockfish 7MB, which is a lightweight version designed for web use, rather than the full version of Stockfish. From memory it is some 60 Elo weaker.

infinitemeditations
Strayaningen wrote:

It is not even close to true that Black loses his queen in all variations, for starters Black's best move is not 16...Qh3 but 16...Qg6, where White is only a little better. But it is true that 11...Qh3 is (thought to be, in current theory) not a very good move and that you have to run Stockfish to like depth 40 before it sees this. This is a well-known theoretical position where the correct move is 11...Qh5 with equality. There is no cloud entry in the Lichess database there, so all you're doing is running Stockfish to depth 28 only. Also it is Stockfish 7MB, which is a lightweight version designed for web use, rather than the full version of Stockfish. From memory it is some 60 Elo weaker.

You seem to really know your openings! Yes 11...Qh5 is the correct move because it prevents queenside castling, which would trap the queen.. 16...Qg6 I looked into it and am memorizing many lines mostly to practice tactics, as realistically no one will play so many master level moves at my level. A very tactical attack that is +6 for white. The queen still can not escape with the mating pressure against the king, and will be lost in all variations, and mate is threatened. It is more losing than 16. ..Qh3 actually. I didn't know that the lichess stockfish is weaker. I need to get the nextmove subscription in the future if I am to be full serious at chess.