Where did i go wrong?

[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]

1. e4 c6 2. f4 d5 3. e5 Bf5 4. Nf3 e6 5. c3 Nh6 6. h3 Be4 7. Ng5 Nf5 8. d3 Ng3
9. dxe4 Nxh1 10. f5 Qb6 11. Qf3 dxe4 12. Nxe4 exf5 13. Qxf5 Be7 14. Bd3 Bh4+ 15.
Kd1 Qg1+ 16. Kc2 Qxg2+ 17. Kb3 O-O 18. Nf6+ *

I analysed this with an Engine. Why is qb6 a blunder? How could i know h6 Was that much stronger of a move? 

Messed up this classical game we played today at my chessclub. Kinda upset. Not because i lost but because i dont understand where i went wrong. I'd love to improve at chess  

its a blunder cz you don't knock a pawn there if they are 2 on the centre with another pawn, you giving the opponent a win /high chances of win

The engine is sugesting. H6 though. It kinda is a 1 move threat. The Queen is binding the oppnents bishop to b2 and holding down the f2 square. I Will admit i didnt think about his Queen going anywhere but e, d or c2. Still struggle to see why Qf3 Was so much stronger than the latter. According to stockfish. My centre is somewhat locked aswell so i thought it would be safe ulike his king. Now after Qf3 i had to take the pawns due to threats on the f7 pawn. Once all that was over stockfish gives him the advantage. Why? Is my knight that bad? Yes its cornered but it also somewhat bindes his Queen to the file.  Also the bishop h4 check was to free my knight so it had purpose.

H6 dont develop anything. Queen b6 does develop the Queen. I guess im Just confused to why the engine gives h6 as a way bigger advantage. 

Well... You don't need an engine to see that white's most (only?) annoying piece is that Ng5, which prompted white to play f4-f5 and attack the kingside, although this plan should be two hundred percent wrong with the rest of his army undeveloped.

Kicking the knight with 10...h6 makes sense, although I would play 10...Be7 myself. I picked this move because I calculated that white's possible attacking moves like f5-f6 lead to nowhere, and I certainly I did not use an engine to tell me that.

This is precisely what you should have done: analysing your game without engine aid. If you go on "analysing" by looking at engine output, you will never improve.

And of course that 10...Qb6 is just one mistake out of many, by both sides.

This is a classic example of how beginners will develop only 1 or 2 pieces and then attack (both players did it).

I see your chess.com ratings are higher than a beginner, so I feel like this was a pretty bad game for your level. As for how to improve on this game, challenge yourself to follow the opening principles more thoroughly.

5...Nh6 is an objectively fine move... and I've played the gurgenidze caro myself (which can feature an h5 Nh6 development in a similar structure) but I feel like telling you that's ok is just setting you up for difficulty, because in most openings it's just a bad way to develop the knight.

So I'd have rather seen you play 5...Bc5 followed by Ne7 and 0-0. It's not as good according to the engine, but it follows principles and is more likely to keep you out of trouble. The point of the opening is not to play good moves per se, it's to reach a middlegame you understand. 

b6 is a good square for the queen, but the position is sharp, so general principles by themselves aren't enough.

For example 10...h6 11.Nf3 Qb6 is winning for black according to the engine.

In other words 10...Qb6 is too slow. White generates play after 11.Qf3

I appreciate the feedback from you all. I don't really trust my own intuition about these positions so it helps to hear your opinions. During the game i felt nervous and  stressed. Early king side attacks puts so much stress on me. I'm afraid i'll miss one detail and have my entire game fall apart early on.

About the position that arises from 11. Qf3 until 13. Qxf5. That's a 5 step calculation which is not so hard to do but evaluating the position after that gets harder. Do you guys continue to calculate after that line or do you evaluate more based on piece activity? 

I get that having more pieces in the center controls more squares but is it always clear to you guys what those pieces will be doing 5-10 moves down the line after calculating the initial first 5 steps ahead? And then there are all the different move orders on top of that. This is something that overwhelms me alot when playing. 

A lot of it is just experience and pattern recognition. I definitely don't calculate 5 moves ahead (10 ply) in most positions.

It's more like after 10...Qb6 11.Qf3 white is making threats and I'm not. White has pieces near my king and my defense of f7 will be is awkward. That's it. Nearly no calculation necessary.

And it is experience... for example I like to step through the game one move at a time to think about what I would play, to try to find the first move I really disagree with. In your game I would have played 5...h5 or 5...c5 (which, somewhat funnily, the engine agrees with)... but these moves break principles so they're not very instructive to mention... which is why I mentioned 5...Bc5

Or to put it another way, some questions don't have useful answers. 10...Qb6 would usually be a good move, it's just specific circumstances that made it bad. Don't worry too much about that. Worry about the things you can control, like playing in a principled way, which means fast development, center control, and king safety. In this game your knight went on an early adventure and your king became unsafe.

Don´t focus only in attack remember you have a king to defend 

I did think of that one yes. Gotta follow development principles more. I tend to overthink things. Clearly his king would Just get away from the checks i made.. 

This is some good advice. I'm going to try some 60min games online to improve. Thanks