How to improve middle game accuracy?

I have figured some ways to get my accuracy up. In the opening I can play for tricks to increase my accuracy because everyone falls for that and a few 90% plus games go very far for myself.  It helps to move knights before bishops and I have learned a law of diminishing returns for pawn moves. In the endgame I have learned to grind (or rather that I should) and it helps, especially in lost positions, because it takes many moves to chase passed pawns and delay inevitable checkmate. It often takes them 10-15 moves to deliver checkmate in 5 and your own moves are mostly forced. Knowing when to resign also seems important to avoid low accuracy moves if you have no idea what to play.

So these are my tricks. I have no clue about the middle game. Any good tricks how to increase accuracy in the middle game? 

To the extend that I could reasonably find better moves when my opponent is not bad. When improving doesn't mean playing more accurately is it not a wasted effort if you can't play a half decent middle game on demand? I think I can blame my opponents if I ever play something good at a critical moment. 

There is nothing at stake, it is just for fun. I'd rather just play a better quality game than win more. The computer evaluation of accuracy is at least objective. 

Any good tricks to actually improve the middle game? 

I have written a comment, at the end of my blog article... 

Good Chess Books for Beginners and Beyond

The comment is actually a brief review of the book....

"The Six Power Moves of Chess" by William G. Karneges

The book is all about how to choose chess moves, primarily in the middlegame.

The comment/book review is in the comments section at the bottom of the web page, immediately following the main text of the blog article

I think the best for kvasir3 would be to forget the term "accuracy". It is not a real chess thing, it is only an invention of the programmers of chess.com. It means nothing.

Hi,

Attempting to have extremely high accuracy in your games is a fleeting idea because humans are not engines, and our play can vary from engine assessments. For example, a human may find an idea to trade down to a completely won endgame, whereas an engine may spot a spectacular sacrifice that ends the game 7 moves later. The engine would penalize the human for missing a win, when the human found a way to win the game easily by trading to the winning endgame position that he knows, rather that spot a complicated and possibly risky idea to sacrifice for a mate in 7. Additionally, some games have positions where the right ideas are easier to find, compared to a complicated tactically charged game when one can easily miss some superior tactical ideas. Accuracy is a loose measure of game performance, yes look at the mistakes the engine points out but take it with a grain of salt; that hypothetical position you found an easily winning endgame by trading down idea is no worse than the complicated sacrifice idea that would have ended the game 7 moves later as long as it led to a force win, and you knew how to do it.

-Jordan