why is my accuracy so low in this game?

Looking at this game

https://www.chess.com/analysis/game/live/6197807440?tab=analysis

it's telling me my accuracy is 14.. (quite a bit lower than my opponent)

It says 8 is an inaccuracy, but I don't see much difference whether I move the knight where I moved it or where it says I should.

There isn't a lot I can do from either place.. where I moved it I attack some squares though can't do anything good from there. And where he says to move I guess I can defend some squares from there, though it blocks my bishop from getting out so prevents me castling that side.

Then his move 8..NG4 I don't see that as such a bad move.. find his piece is hanging which isn't so good but he's kind of goading me to attack and if I did i'd lose my queen. His piece isn't under attack, and could cause me some nuisance.

His move 9.. knight forking me. Leads him to trade his knight for my bishop and he prevents me castling. Though the engine says his move was an inaccuracy and he should have kept his knight back.. Engine tells me i'm meant to trade knights. But how am I meant to know i'm meant to trade knights.   There is a philosophy(not shared by Ben Finegold) that when way ahead on material, to trade pieces, but I'm not way ahead on material.

14. I screwed up I missed an opportunity to check him and take his rook.  It does say that'd have been the best move. But it says I still did a good move there. so it didn't call that an inaccuracy.

15 was a tiny screwup that it calls a blunder. It's telling me I should've moved my queen to G5. And I can see that'd have protected  a pawn. But He could have done h6 and attacked my queen while opening a bit of room for his king to escape if need be.  Sometimes it's good when his king has hardly anywhere to escape.

16 it tells me I should have moved my queen one spot down, I don't know why.

17 it tells me to move Qg5  but he could reply with f6 , i'm really not sure why the engine is so keen on me being there.

18  it's telling me I should have moved my queen to G5,.   I don't know why.

24,25,26 were a bit crazy.  and I can see how the engine would call that inaccurate or blunder or mistake.  I kind of feared i'd lose a rook if I wasn't careful, and that if that happens then i'd probably lose the game.. I spotted he had a one move checkmate I wasn't sure how to block it and feared losing my rook..I tried to bait him into moving his bishop. And then I realised that I could mate him in one. He failed to mate me in one maybe 'cos he was so keen to take my rook.  And I mated him in one. I can see how an engine would not look fondly at that.

I don't understand why it thinks I played so badly.. all the many times it thinks I did inaccurate moves or mistakes. 

I don't think either of us were really leaving any pieces hanging that got captured(pawns aside). Apart from one case and that was him losing a bishop.  Not me. Neither of us blundered any important pieces i.e. none of us blundered any pieces of more value than a pawn.  No "doh" moments(apart from him missing his mate in one). 

Maybe the computer would have preferred my play had I castled, but me ending up not castling,  was actually thanks to his fork move 9, which forced me to use my rook. to kill his knight on move 11.   And the compute didn't like his move 9 anyway   

I thought we both played well.. apart from 24-26 at the end where he missed a mate in one and I failed to defend the mate in one, but  I did some strange antics and beat him.  And he was less good than me in that exchange!

The engine evaluation in general punishes for the mate-in-one blunder a lot more (if you look at the move-by-move difference, it gives you a score of -990 points for that move) than from just blundering a piece. In double-edged blitz games the accuracies are lower than in games, where the positions are simple all the time. In general the accuracy cannot be interpreted in a straightforward way to how well you played, it depends on how complicated the game is and how good your opponent is.

It is a bit hard to tell whether you (the opening poster) are trolling or talking seriously. Pieces are hanging left and right, both of you allow mate-in-1 and you complain about "inaccuracies"?

Where are my pieces hanging left right and centre? Can you give some examples?

All I see are

7 my knight is hanging but that is because I am trading knights
14 a hanging pawn as a result of killing a piece
21 I see a hanging pawn on h2 

Thanks

10. Qd2 instead of winning a pawn on e3 is just weird.

13. - Qh3 loses a rook (Nc7+) but you don't take it.

There are also a few moves like 5. Bf1 or 15. Qb4 that are very difficult to explain.

I won't analyze this because others seem to be doing that already. As for the accuracy score (CAPS), I really like it on chess.com, but most players simply do not understand what it actually is. The game "accuracy" is your moves (and opponent moves) compared to an engine analysis (chess.com uses Stockfish). If you played the exact moves the computer recommended on each move, then you score 100%. This is NOT the same as saying the "perfect move" because it depends on many things including the engine depth. If you run your game through a different depth, then the accuracy will change even though it is the same game! (Longer analysis or deeper depth is probably more reliable). 

The accuracy also doesn't account for many other things such as "book moves", how "difficult" those moves are to find (I play like a computer whenever the line is forcing/I have only one legal move xD) and also doesn't account how easy the position is to play. For example, let us say your opponent blunders the Queen on move 5 and you take it for free. Now, for the rest of the game, it is super easy for you to find good moves because you are already winning by so much - practically any move you play would be winning - so you might get a high accuracy without much effort.

IM @DanielRensch explains it better than I could, so I'll let him explain it for me

Re 13 Qh3

The previous move by my opponent was 12...Qh4+

So on the next move, 13, I can't do Nc7+  at that time, because I am in check and I have to deal with that!

Re 5 Bf1 it's not strange at all. My Bishop is under attack from his knight.  Even the computer says the move is "excellent".

15 Qb4 I know why I did it, but it was a blunder,  the computer's suggestion and I am not  arguing in favour of that move.  And it was a very small blunder, it cost me a pawn.,  It has an advantage that if he were to move his queen anywhere then in one move , to B5, I could put him in check.  Then he' d have to move his king and he  then wouldn't ever be able to castle. But i'm not in favour of that move. that's one blunder that I mentioned in my first post that I understood. 

Re "It says 8 is an inaccuracy, but I don't see much difference whether I move the knight where I moved it or where it says I should.": 8. Nd5 Nxd5 9. exd5 Qa5+ 10. Bd2 Qxd5. You lose a pawn and the opponent gets a big center.

Re "Then his move 8..NG4 I don't see that as such a bad move.. find his piece is hanging which isn't so good but he's kind of goading me to attack and if I did i'd lose my queen. His piece isn't under attack, and could cause me some nuisance.": 8...Ng4 just misses the line I gave above, missing a huge advantage. Also after 9. Be2 that Knight just has to retreat. Black lost time and just helped White to develop.

Re "His move 9.. knight forking me. Leads him to trade his knight for my bishop and he prevents me castling.": No, after 10. Nxe3 dxe3 11. Bxe3 Black just lost a pawn and traded his only developed piece, also helping you develop your Bishop to an excellent square.

Re "Engine tells me i'm meant to trade knights. But how am I meant to know i'm meant to trade knights." The knight trade itself is not the point. The point is that it leads to White winning a pawn and develop a Bishop versus losing the right to caste (and a pair of minor pieces is traded anyway).

Re "14. I screwed up I missed an opportunity to check him and take his rook. It does say that'd have been the best move. But it says I still did a good move there. so it didn't call that an inaccuracy.": Yes, you missed a chance to win material. Btw note that what the engine classifies as "good" has quite a wide range from actually quite good to significantly worse than other moves.

Re "15 was a tiny screwup that it calls a blunder. It's telling me I should've moved my queen to G5. And I can see that'd have protected a pawn. But He could have done h6 and attacked my queen while opening a bit of room for his king to escape if need be. Sometimes it's good when his king has hardly anywhere to escape.": Tiny screwup? You went from totally winning to just worse in one move! Black will win your e5 pawn, undouble his pawns, protect his d4 pawn getting a nice center, attack your Queen and prepare to develop his Bishop and castle, all in one move.

Re "16 it tells me I should have moved my queen one spot down, I don't know why.": On b3 your Queen protects d5 and does not block your undeveloped Bishop.

Re "17 it tells me to move Qg5 but he could reply with f6 , i'm really not sure why the engine is so keen on me being there." 17. Qg5 f6? 18. Rxf6 gxf6 19. Nxf6 family fork. And the engine wants your Queen to get there because it wants you to attack that d..n King!

Re "15 Qb4 I know why I did it, but it was a blunder, the computer's suggestion and I am not arguing in favour of that move. And it was a very small blunder, it cost me a pawn". No, no, no. Look at my previous post. That was not a small blunder.

Also note that 15. Qg5 not only brings your Queen to an active position on the kingside and protects your pawn, it also is very annoying for Black, because now ...Be7 runs into Qxg7. If Black plays ...h7, this h7 pawn may become a target later (e.g. Black castles, White's Queen is on the g file, then the g pawn is pinned and Bxh7 is possible. Sometime also a Bxh7 sac might be possible). You want Black to move the pawns in front of the King because as long as they don't, they maintain maximum flexibility. Every movement is a commitment and creates weaknesses.

Re "But I see I guess it's also two pawns in the same file even if they are not adjacent?":

Yes, it does not matter if they are adjacent. What matters is that these are doubled isolated pawns, which is really a bad thing, because none of them can be defended by a pawn. But after you allowed dxe5, suddenly Blacks pawn structure is perfectly fine!

Yes, there is no pawn on e3, but you can play 10. Nxe3 dxe3 11. Bxe3. This wins a pawn. This is what the engine suggests. If you want to preserve your knight and give the bishop instead, play 10. Bxe3 dxe3 11. Nxe3 - that also wins a pawn although that is not as good. Instead you choose to NOT win a pawn but leave that monster knight on e3, which is absurd. To top all this, next move your opponent chooses to give up that monster knight for the undeveloped lame f1 bishop (an even weirder decision). These are not inaccuracies, these are blunders, practically equivalent of hanging a full piece each move.