is studying modern masters useful?

It's a well common advice to study great player's games.

But I was thinking: with the amount of theoretical preparation of modern GMs (they even have whole teams of theoreticians!) honestly what's the point in seeing what they did?

Their decisions up until move 20+ is based on nothing else than home preparation. So they do moves not because it's logical or has a "human" strategy, but because an egine said +.05 or something.

Also, certain lines might be chosen in attempt to avoid the opponent's preparation (e.g. nobody would play the Sveshnikov again Gelfand because he knows like 40 moves of theory).

Another consideration is, as a GM expects a "professional" response, they don't particularly focus on the lines where the opposition can go wrong. In amateurs games it is very important to give the opponent the chance of messing up.

Now, a particular game might be used to illustrate a theme or a strategy, so of course I study games in strategy or opening books.

But picking a random game from a modern GM and just looking at it on your own, I honestly don't think it's very useful.

I think you make great points. I enjoy going over master games, but I feel like I understand more (who knows, I might be kidding myself there) and get more from studying older games from earlier eras. There are a million exceptions and of course if you are studying a particular opening the more contemporary the better -- but contemporary super GM chess games often feel like an unpleasant combination of inscrutable and dry (to a patzer like me.) 

I understand your point. At our level we should be studying games by older GM. Capablanca for example. Take a game and annotate every move. show all the candidate moves you considered. Then go over the game with the computer and see if there are moves you liked that have a rating of more than 1/2 a pawn off the best move. Those are the moves you need to concentrate on.

Between Capa and the 21th century games, you can also benefit from the post-war masters (1950-1970) and the modern ideas developed in the pre-computer era (1970-1985).

Unless you want to be completely up-to-date with current theory, that will probably prove more useful in the long term.

It can be, especially rapid games, but generally older master games have more flavor to them. 

The sentence highlighted in bold red type from your post summarizes the points that you made in your post very well.

The key factor and the one you did not mention designing an opening repertoire which reqires building an opening tree which today the computer builds for you.  The hard work begins, and requires about 2-3 years, to become proficient and competent with your opening repertoire.  This is the point where you become selective about what modern GM, IM games to study.   

What I plan to do is go over the games of Carlsen from when he was an expert to about 2500 strength.

There`s a lot to be said for studying master games in more-or-less chronological sequence... starting with Greco, Philidor and their contemporaries, then Anderssen, MacDonnell and LaBourdonnais, then Staunton, Morphy and Paulsen, and on through Steinitz, Lasker, Capablanca, Rubenstein, and so on.

There is a general rule of thumb.  The lower your rating, the older the master you should be studying.

Does this means it's "bad" for a 2200 player to go over the games of Capablanca?  Of course not.  However, the opposite should be avoided.  It won't do a 1200 player a whole lot of good to try and understand Kasparov's games.

Think about the fact that a lower rated player should be learning opening concepts and NOT opening theory along with Middlegames and Endgames.  Opening theory was not nearly as well known in the 1920s.  You didn't have computers, databases, etc.  You were on your own to do the research.  In the computer era, and the introduction of the chess engine, analysis has been done much further in depth.  A 1200 player should not get involved in this for many reasons, including, but not limited to:

• A 1200 player is going to play another 1200 player.  Even if the first 1200 player studied opening theory like a 2200 player would, the odds that your opponent remains in "book" for you to execute this "new" novelty that Joe Schmo finds in the Najdorf Sicilian on the 16th move are literally ZERO!
• If said 1200 player in the first bullet knows this "novelty" all the way out to move 16, does he really "know it"?  Or is he just merely capable of parroting the moves?  Your opponent is going to deviate!  What good is it to be able to parrot long lines of opening theory if you have no clue why you are making the moves, and hence have no idea what to do when your opponent deviates?
• The use of computers and databases has now created a lot of "exceptions".  Knights on the rim are dim, EXCEPT in such-and-such a line where because of some 13 move combo, the knight belongs on the rim!  Older players that played via opening concepts didn't have all this "trash" to keep in their heads.  Stick with the general principles until you reach at least 1800.

Also, to tie together with the 3 bullets, think about what older players are known for versus more modern players.  With Nimzovich, he's extremely well known for the concepts of overprotection and blockade, two VERY IMPORTANT middlegame concepts!  Kasparov?  Yeah, he was really good, but he made no "new" contributions to chess outside of opening theory.  With Kasparov, it was all about changing the assessment of opening lines, like maybe a line that was once thought to be good for White is now a draw because of some magical move Kasparov found in opening theory on move 17 that has changed the entire assessment of whatever said opening you are dealing with.

Their moves are more than the average player can comprehend so if you try using some of their tactics you might fail because you wont understand what to do when slight changes in the game equation arise.

Most people that come to this site are looking to become stronger players within a reasonable amount of time.  A time frame of 2 to 3 years to reach an OTB USCF or FIDE rating between 1800 - 2000. 

In the days before computers it was necessary to build an opening tree for your selected opening repertoire by hand.  It would take an average of 8 - 10 years to build an opening tree by hand.  It required combing through thousands of games, old and new, one at a time.  Today with ChessBase a computer can build an opening tree for you based on your selection of openings (4-6 openings) that comprise your opening repertoire in a relatively short time.  Then you must practice your selected openings 2-3 years in order to become proficient and competent at playing those selected openings from the White side as well as the Black side.

To become a strong player requires, the most difficult part which is, building and becoming competent at a selected opening repertoire.  But there are 4 other visualization pattern memory banks that you must build into your brain.

The 5 visualization pattern memory banks that you must build are:

1.Tactics visualization pattern memory bank

2.Matng Net visualization pattern memory bank

3.Endgame visualization pattern memory bank

4.Opening visualization pattern memory bank

5.Middlegame visualization pattern memory bank 

The experience as a strong player after having acquired the 5 memory banks is as follows:  When you look at a position to analyze it the move jumps up off of the chessboard and smacks you on the forehead in a flash!!!

To summarize, to become a strong player takes about 2-3 years.  The process involved in getting there within the 2-3 year time frame is acquiring the 5 visualization pattern memory banks, and playing 60-72 OTB USCF or FIDE rated tournament games per year.  Also, alot of skittles games and some blitz games to reinforce the 5 memory banks and practice with handling the clock under time pressure. 

If you have any questions or would like to know more please let me know.

My question regarding this topic is how to study an annotated game. For example, I got a book by Botvinnik with a lot of his great games, annotated, and what I am doing is setting a board and follow the text. Moving what it says and reading the annotations when there are. I try to undertand the different lines it analyses but I go through a game quite quickly. I dont think I am doing this completly right. What do I have to change?

After the opening, should I not look at the next move and try to figure it out by myself? 

Or maybe follow the text and just try to understand the threats and consequences?

Thanks in advance!

Im gonna be the "weirdo" of this thread.  As a USCF A/B player (Depending on how im playing) I follow the advice that we are all told.  Play up a section, so I study Expert/Master level games.  They are easier to understand, but still a challenge. 

The modern day games teach me very little, if anything. I suppose it's because modern play is too complicated for me to understand. But the older games, perhaps because they are not as sophisticated as today's games, have taught me quite a bit. Plus in my opinion they are a thing of beauty.

If i do study a modern GM's game.  I go right past the studied to death, computer aided opening, and get right to the good stuff, the middlegame.

I think Capablanca and Rubinstein are particularly good choices for weak players to study. They had a style of elegant simplicity that makes it easy to follow their ideas. A typical Capablanca game leaves me thinking "How simple. How obvious. But how does he do it?"

But...how do you study annotated games? Which is the way?

Judit Polgar has a three book series that's said to be really good.  She's retired but might be current enough.  Ivanchuk also wrote a game collection, as did Kramnik but he's way overdue for updating it since he's played many fantastic games since he last wrote it. 

No FM or GM plays a move because it says +0.05 on a computer, not unless the very unlikely scenario they had zero preference either way and it was a total toss up. I don't know how you got that idea. They use computers to ASSIST in SOME specific openings, especially those involving hard complications. For example the Botvinnik Semi-Slav. In moves where principles are more important they often go against the computer by 0.3 or more.   

I certainly agree though that the people under 2100 who say "of course, I can't play that anymore because Nakamura came up with 25. Nh4 last November" are deluded.