Why is this position so strong for Black?

Sort:
mkp5433
I encountered this position in a recent game that I played. Though I ended up winning, The computer analysis really surprised me in this position, showing an evaluation somewhere in the -6.3~-6.7 range depending on the engine.
However, Though I agree that black is winning, I don't understand why the engine favors black to the extent of two minor pieces, or a full rook advantage.
I recognize that black has more King Safety, a decent pawn structure, and good development.
But White doesn't have a bad pawn structure, White's King is not unsafe, material is even, and there appear to be no tactics (Though I could be missing something).
There are also not very many imbalances to take advantage of.
The engine suggests Qc7 as the best move, and I understand why this is a good move, (doubling up on the semi-open file, preparing to attack the king).
However, In the game, I chose NxBd2, and this is recognized as an inaccuracy by the engine, dropping the evaluation by almost 3 whole points. I saw this as a good move, by removing a bishop for a knight, generally a good principle.
In general, Why is this position rated so strongly? Why was my move an inaccuracy? What strategy should I have in this position?
For reference, This game was played at about a 14-1500 level on Lichess.
First time posting, so thanks for being kind!

TheDishesAreDone

White's King is completely unsafe. It has advanced to the second rank with effectively no pawn cover while the Queens, Rooks and most of the minor pieces are still in the game. It's a position you would expect to see the King in during the endgame, not the middlegame. Whereas, Black's King is about as safe as it could possibly be.

White's e3, d5 and b4 pawns are all targets.

Black can immediately take control of the open c-file with devastating effect.

For instance, after playing something like 1... Nxe3+ 2. Bxe3 (forced due to forking King, Queen and Rook) Rxc3 3. Bxb6 Qc8, White just has too many positional weaknesses to contend with, particularly the exposed King. The engine's 1... Qc7 recommendation looks like a more effective positional squeeze, with one potential follow up being Qb7, targeting the Rook on a6 and the d5 pawn sitting on the same diagonal as White’s King. Also note the Nxd2 move played in the game is a much stronger threat after the Queen has moved to c7 first.

 

WBillH

"material is even"

Not for long.  White has a lot of hanging pieces, so it's not a matter of finding a good tactic for black, it's a matter of finding which one beats the rest.  You have to go pretty far down the list of candidate moves to find one that isn't winning.

Nxe3+ immediately wins a couple pawns and creates a passed pawn for black. 

Wildekaart

There is little harmony between White's pieces, and Black can make all sorts of discoveries and attacks to take advantage of that. To say it's a difference of -6 I too think is pretty exaggerated. It can say it's -6 but there are a bunch of examples where one side is down by that much according to engine evaluation and definitely not out of the game.

This is a fun line to walk:

As Bill said, it's hard to find a move that doesn't win. Unless you continue to screw yourself over for a few more moves.