Well one of my approaches is to think if you never understand the reasoning behind a good move you'll never play it yourself. So when you come to one of those "I don't really understand this move" in a GM game, don't skip passed it. Try and find about it and the ideas behind it, and experiment with the idea in some of your own games.
do you use a board or just read the game blindfold
The guess-the-move approach can be useful if you are at least at an intermediate level.
I'd rather just stop at all key points where I just don't understand a damn thing and save those positions and bug stronger players I know to explain it out to me. (A luxury I have but most players complain that they don't, I know!)
If you have an objective mind and actually know how to use an engine to analyze some of these "WHY" positions, you might get lucky and start seeing see analytical (forcing reasons) why Move X was played or why Move Y was NOT played but it is a crapshoot if you expect to "solo" study a non-annotated game and gain insight about strategic/positional nuances. Engines spit out eval. scores, not well reasoned sentences explaining things clearly the way a human does.
Instructional annotated game books would be a safer bet.
Do most of you guys use a computer or a chessboard when playing over games?
Take the board. Seriously. 2d makes chess one step more abstract than at a real board, *and* using an engine will pretty much make you think and learn *less*. You will be "observing" rather than being involved. Trust me on this, get the board up. Move the pieces. Frown and ask yourself why this move or that move was made, instead of staring at an eval and a "principal variation". I think you'll learn tons more that way.
Take the board. Seriously. 2d makes chess one step more abstract than at a real board, *and* using an engine will pretty much make you think and learn *less*. You will be "observing" rather than being involved. Trust me on this, get the board up. Move the pieces. Frown and ask yourself why this move or that move was made, instead of staring at an eval and a "principal variation". I think you'll learn tons more that way.
+1.
Socratic "thinking" trumps "1-way transmission" from a book or engine evaluation to your brain. You engage your brain instead of just "sitting there".
Just got my copy of Zurich 1953 by bronstein..... usually I would sit at the chessboard at play along... but im considering playing through the game at my computer with the engine now... trying to do the whole "guess the next move, see how it compares to the actual move, see if the engine likes your move better, the text move, or different move"
this process DOES take a lot longer though, but I feel I understand the games better
Also.... how do you guys go over unannotated GM games that you download off somewhere like chessgames?