I am currently rated 1850. My coach tells me that in order to get better at my game, I should ''analyze grandmaster' games''. On the surface, I understand that he wants me to learn from great players' moves. But I'm unsure about what I should be ''analyzing''.

In particular, I have the following problems:

  • When I look at GMs moves, I see GMs responses. So it's difficult for me to learn why the GM 1 played this particular move, when the GM 2 plays an equally "perfect" move. As a 1800 player, I cannot see the +0.1 advantage that the GM 1 just obtained by his move.
  • If I put an engine like Stockfish on the move, I will have a bunch of move opportunities, but none of them ''speaks to me''. Why is one move slightly better than another? Stockfish will not tell me.
  • Finally, like Kasparov said, ''Life imitates Chess''. So in principle, I should make each move with an underlying grand plan. But what is the grand plan? I do not know.

So how can I ''analyze grandmaster games'', so that I can learn from them?

Please give me some guidance.


I agree with you, more than your coach. Just analyzing a GM game, even with a computer, is not helpful just by itself if you have no idea what is going on.

For example, just earlier today, I answered a question here, where the person gave a computer line that made no sense as far as the plans for the position were concerned, so the computer was really no help at all there.

I think what you need are grandmaster games that are WELL-ANNOTATED, to tell you what is going on in them IN WORDS, not just variations, although concrete lines are clearly important too. Another thing that happens in modern chess is that games between IMs/GMs have become so advanced that the battles between each side's plans have become too complex for learning early in a chess "career" because the plans are so effectively thwarted by the opponent it is not always clear what each side was aiming for.

What is much better are the games of old masters like Capablanca and Alekhine because their opponents, while not slouches, were not on their levels; so you can see plans more clearly executed from beginning to end as the plans were often not met with the type of resistance that you would see today at the top levels.

I would get collections of their games, written by them, and play through those games. I have also written here extensively about opening pawn structures, and as you read any game trying to understand it, you need to have some understanding of what you are striving for in any given opening.

Another book that is extremely good for learning is "Learn from the Legends: Chess Champions at Their Best" by GM Mikhail Marin.

Learning to calculate is really a different skill, but one that is reliant upon how well you understand chess (see above), in general, and what you are aiming for (that is, what is the plan).

Calculation prowess can be gained by combining two practice activities: Reading pure tactics books, and reading books that have non-tactical positions for you to practice calculating the best move. Also, try practice reading annotated games without a board, including the annotations, to work on vision in your head.

Tactics books are self-evident as there are a million of them out there. Some excellent books that will give you positions to calculate (mostly) non-tactical positions include:

"Positional Chess Handbook: 495 Instructive Positions from Grandmaster Games" by trainer FM Israel Gelfer. I cannot even begin to tell you how much I love this book. It is incredible. (Do not get the Kindle version as there are way too many mistakes)

"The Best Move" by GMs Hort and Jansa.

"Test Your Chess Skills: Practical Decisions in Critical Moments" by GMs Sarhan Guliev and Logman Guliev

These will help you improve your pure calculation.

  • 1
    "Just analyzing a GM game, even with a computer" I'd say "specially not with a computer" – David Mar 31 '20 at 10:07

So forget about Stockfish, forget about Kasparov, forget about GMs.

You are 1850. You understand something about the game. Not everything will be correct, some of it will be.

Start analyzing. It can be your own games, it can be any old game, it can be a random diagram you saw in the newspaper.

Set up the position, and try to figure out what is going on. What are both players trying to do. Are there any tactics, threats? What are the candidate moves. And finally, what is the best move. Write it down with some short variations as arguments.

Then see what was actually played. If it's the same move, fine, move on. If it's different, was it at least one of your candidates? Why didn't you choose it? If it wasn't, do you understand why it was played, does it fit in with the "story" of the position you had? If not, try to figure it out for a few minutes. If you can't, shrug and move on.

If you can do this with an annotated game, you can compare your annotations with the book's afterwards. You'll have missed mostly everything! But it'll get better.

This is very hard work, but I think everybody would improve doing this regularly.


I’d also recommend Silman’s The Amateur’s mind — I’m currently rated a bit above 1800 and I found it much more helpful for someone to show me how think about how I think about a position and why do I get some things often wrong. It’s really interesting to see yourself in someone else’s words when they’re asked about a position.

In short, I think it’s good to think about why/what did you get wrong instead of what did a GM get right.


More than analyze you should play over hundreds of them in the openings that you specialize in.

Then you get an intuitive feel for the moves to make and how the game continues into the middle game.

Just analyzing for the sake of analyzing is just busy work if you are not learning to play better.

One old book said to increase your rating then you need to stop playing for a year and do some useful studying. If analyzing helps you study then do it, but for me I find just playing over games and occasionally looking at the annotations for a given move when I would have done something else is more useful.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.