I was looking through some games before they get wiped for live 2 and I ended up clicking the analyze button by accident. So i waited and when it was done I took a look. After going through it, it seems like it gets really harsh with the annotations. I won but I missed some tactics.

the computer analysis here is always like that, every 3rd move is a mistake if not, then an inaccuracy especially when you win.

I guess the computer like Rb1 because it discourages Nb4 by black, which would cause white some discomfort. And Nb6! is such a computer move. It looks un-natural but man is it strong.

no you shouldn't worry about those inaccuracies you were playing fine

probably, the computer liked Rb1 so much because it 1) takes control of the b-file 2) develops a piece and 3) gets the rook in position to move out. In addition to preventing Nb4 mentioned earlier, the black bishop is hindered unless it wants to go back to starting position (which gives up a tempo) or exchange bishops (which will stick the black knight in a not good spot) and give up a pawn.

This is my thinking, but I'm not so good. I could be wrong on one or more points here.

hmm i thought you played really well!

Thanks for taking a look. I'd never seen that many question marks in one game before so I was nervous.

Well, computers are programmed to calculate tactics and always go by general principles.  Because it couldn't see anything tactical, it liked the idea of putting the rook on the open file, an idea it was programmed to recognize.  Sometimes, computers make horrible blunders in the endgame because it cannot calculate tactically all the way to a gain of material or promotion.  So it makes a move that gains space and loses.