Question on why nc6 is so weak according to stockfish

nc6 just isn't the right way to coordinate your pieces, the knight if it was put on d7 would have had a lot more prospects with some future Nc5 or Nb6-Nc4, putting pressure on all of White's camp, he'd probably have to give up the bishop pair

Yeah in this position the knight's potential movements after Nd7 are more promising than Nc6. That being said, these aren't really the kind of mistakes you have to worry about. Worry more about finding moves like Kh7 or bringing that rook over to cover the g-file.

I see thanks

The reason is both positional and concrete: in these English Attack positions, Black wants to meet Nd5 with ...Nxd5 exd5 Bf5, immediately neutralising the diagonal White's bishop most wants after exd5 happens.

However, with the knight on c6, ...Nxd5 exd5 forks knight and bishop. Because of this, Black will be forced to do something suboptimal, like taking with the bishop which allows a powerful unopposed bishop on d3, or not taking at all, which leads into the positional drawbacks of ...Nc6:

The c-file is blocked, nothing is being done to support ...d5, and there are no good maneuvers (or manoeuvres for you) for the knight

Kh7 wasn't the best move. It worked because white played Bxh6.

I wasn't anticipating bxh6, I thought white was going to go g5

Thank you, I was trying to play bxb3 and then somehow get the knight to d4

I would not bother so much about Nbd7 being better or worse than Nc6.

Nbd7 is the better move, as whenever white plays Nd5 you can chop it and then try to attack the d5 pawn by ...Nb6 etc- but this is not critical.

But I would surely play 12...0-0 in this position only when being totally drunk: White has a killing attack on autopilot, and 13...Kh7 does not change this fact at all. As you said before, 14.g5 or 14.Nd5 or even the typical 14.Kb1 prophylaxis first (white has no reason to hurry up), and it's very unlikely that your king will survive for long.

Yeah, castling into the potential attack wasn't very smart, what would you recommend instead of 12.0-0?

One idea is to play 12...h5, which slows white's kingside offence and postpones kingside castling for a while (and anyway, generally speaking it's safer to castle kingside in such positions with a pawn on h5 rather than h6). However, after the typical 13.Nd5, it's already apparent that the knight shouldn't be at c6 as there's no good way to capture at d5.

So, I would rather play something like 12...Na5 (and possibly ...Nc4 after that) and wait for white to show his cards. White's game is easier, but I think this is no biggie.

Not checked these lines with an engine for complex tactical shots, but I think (hope?) they are OK. 

So something like this?

regarding the h5 variation

No I suggested 12...h5 (after 11...Nc6 12.0-0-0), but I think 12...Na5 is more to the point.

And in the opening, your 6...e5 is a mistake - unlike the similar line in the English attack the bishop is not on e3 but rather g5, the f6 knight is pinned and so white can well play 7.Nf5 when you are in trouble, already.

The real answer to your question is because your engine source is very weak compared to others.

This will hurt your game more than anything else.

Best example I saw was your move 13...Kh7 (the brain/best move/2 minute calculation). How was that even a "good" move?

You might use the site analysis/game report for a blunder checker...but anything more will probably hurt more than help you.