Calwhlating to the end

I was just going through Andy Soltis' excellent book 'The inner game of chess'. He presents an extremely beautiful position from Reshevsky-Botvinnik. At first I looked at the position and thought it was one of those messy positions with no clear answer. After reading the analysis, I realized that it was actually a deeply beautiful position. Here botvinnik made the mistake of losing one tempo in a rook maneuvre. He could have made use of a tactic to save a tempo, whereby the opponents pawn would be one square behind and would have helped black to win. However, with the loss, it was white who won.

In a real game, I find that I cannot think in this way. The difference is that masters calculate to the very end and know exactly the key fearures of a position. This was an endgame position so I was not too disheartened since I really like endgames and I am never ashamed to learn a thing or two when I get such a chance.

However, my main problem is the middlegame and especially the 10th move or so when the opening is just finished and I am at a total loss at figuring out what to do.

My question is, how much should you calculate. Would you calculate to the very end or just try and improve your pieces or what to do??

I calculate until my head gets fuzzy. Improving my pieces takes calculation too, so calculating as far as I care to go, or I am able to go, is really the whole point of chess, isn't it?