What are the different approaches you guys use to analyze top-level matches and learn from them?

I've only recently started backing off my number of games played versus games studied, due to the immense amount of reading I've been doing lately.  This led me to ponder what the different approaches are for truly learning from top-level matches.

I've tried loading the PGN into a computer program (in my case, Shredder 11) and either running a full game analysis on it and reviewing the recommended variants or stepping through the match, one move at a time with live analysis running in parallel. 

I've also tried simply walking through matches with any random PGN viewer, and after the first couple moves, I start to list candidate moves for both black and white and pick my favorite candidate move before moving forward to see what the player chose.  I kind of adopted this approach from the various Silman books I've read. 

So far, I'd have to say the candidate move approach without computer assistance seems to be far more pleasurable and instructive for me personally, as it really forces me to think.  I'm debating if it might be an interesting study to keep track of how many times I guess the correct move for a match and see how much my accuracy increases over time.  So for example, If I only guess 6 out of 60 total moves, then I have a 10% accuracy base to propel upwards from.

Has anyone tried keeping track of these kinds of statistics for themself?  I think it would be very very interesting!  This approach actually makes me ponder something that Chess.com might want to loosely think about (shares some similarities to Vote Chess) - a way of proposing candidate moves for past matches and keeping track of your accuracy in choosing the move that actually happened.  You could possibly think of this as a PGN Game Study Aid :). 

Erik, if you're listening to this, I'd be interested in getting your feedback.

Trying to guess the moves for one side is called Solitaire Chess. Pandolfini wrote a book on it, and publishes a new Solitaire chess game in every issue of Chess Life. It is a fun and useful way to study, especially when annotations by a stronger player are available so you know if and how you went wrong.

There are some dangers to going it along and scoring that way:

1. In many positions there's more than one correct move, and how many 'Karpov' or 'Tal' moves you find may reflect your style more than your skill level.

2. Suppose you find 9 good strategic moves, but then you chose a move that hangs a piece to a 3-move combination. Do you deserve a 90% score?

When I seriously study a game I do it differently. I assess the position, examine a few candidates, and choose what move I would play. After writing down my analysis for all the moves, I compare them to annotations, my database, and an engine. I don't score myself; I'm focused on learning whatever I can. I choose games based on my opening repertoire.

I don't do such rigorous studying anymore as mentioned above, but I agree that the computer-generated lines shouldn't be considered study. The only way to analyze your own games is to first go through them yourself, and only then to utilize the computer.

What I do often is to look at a position to see if I evaluate it properly - what are the key lines, what are the key pieces, what are the themes which guide the play for the next several moves. That's a weak sort of candidate move study, but it is useful nonetheless and can be done without a board. ;-)

I've always followed the advice David Bronstein gives in his book 'The sorcerer's apprentice'.Quoting Bronstein: "First,play through the whole game without hesitating more than a couple of seconds at each move.If you have the urge to pause longer - don't!Just make a mark in pencil and continue to play the game to the end.Then put the book aside,get a cup of tea or coffee and try your best to recall from memory the spectacle you've just seen.Try to establish hte reasons why decisions were made.Second,play through the game again,somewhat slower this time,and mark in pencil everything you didn't see the first time.Third,now go straight to those pencil marks and give your imaginative and creative energy free reign.Try to play better than my opponent and I."

A bit further he says: "If during stage 1 you made no pencil marks at all,don't look at this game again.Go on to the next one that,hopefully,will give you more pleasure and satisfaction."

 Of course,in this computerage a fourth stage can be added.After I went through the first 3 stages I check the moves with an engine to see if any tactics were missed.

It's quite a bit of work but the rewards are great.