Usually you will calculate in a critical position. For example, all your pieces are ready for a breakthrogh, or you have a possibility to exchange some pieces.
What you should look for are
1.Possible forcing countinuations-- look for check, capture, attacking moves.
2.Anticipate opponent's possible ideas that might counter yours.
3.Make sure it works; you didn't blunder something.
After that, you can compare to find the best line. To perfect this, it take other skills like evaluation skill or positional judgement, which are even more important than calculation skill.
Notes:
- A calculation should not take more than 20 minutes
- Don't look too far. looking 5-6 moves ahead is good enough. Judging the position is more important.
- In a normal position, you don't calculate at all, it's more about intuition.
For a brief period of time I didn't play games as much on this site and instead did tactics trainer and some of the chess mentor lessons. I have noticeably improved at tactics trainer and calculating in general but this has not lent itself to better play. I was regularly playing at an above 1800 rapid rating but now I am mostly in the 1700s. I calculate through any forcing variations but often it doesn't directly lead to a winning position. Also I am more likely to miss simple moves for my opponent that lose me the game, whereas I feel that I used to have a better awareness of all the immediate playable moves. So my question is when do I calculate and what should I try to calculate? I know I have to play general moves to improve my position in relation to my opponent's position when there are no forcing moves, but to determine which general moves are best calculating some variations is still required. However, too much calculation wastes a lot of time and at times mentally removes me from the actual position to the extent that I make careless mistakes.