What do you think is the most effective way to improve your game?

I'm particularly interested how players rated 1900 or above would answer this question.  In your opinion, which form (or forms) of training or practice or study give the most bang for the buck?

(Caution: I have dared to answer your post, though my USCF rating is but 1540.)

The single most valueable activity to achieve improvement is the clear identification and correction of one's own weaknesses. Well, yes, that sounds like mom and apple pie, but it is the truth. A ruthless examination of your losses (particularly) that ends with a determination of the reasons why you lost and a program of learning/skill development to remove the causes is the best medicine.

Here's some examples of what I am talking about.

If an examination of my losses reveals that I'm missing my opponent's threats [I know this by measuring how surprised I was by his/her move(s) and plans], then I know where to put the work in (I must examine not only my moves but my opponents responses, I must not only understand what my plan is but try to identify what my opponent is trying to do, etc.).

If an examination reveals that I am making bad pawn moves (I can determine this by engine analysis), then I must find the reasons. Do I not know what the typical/ideal pawn structures in my opening systems are? Do I not examine all my opponents options when I offer a pawn capture? Do I not appreciate what weaknesses I am creating by my pawn moves? Do I not see what outposts my opponents' have if I make certain pawn moves? (And there are surely other possibilities.)

Sometimes we can easily figure out what we're doing wrong, but have a much more difficult time trying to find out WHY we are doing it that way. Is it because I am missing an important piece of information? Or is it because I have too little experience with the positions from which I went wrong? It depends, so some self-examination is required. Getting the counsel of a strong other will help, but that person cannot do all the work of identifying why we are making OUR mistakes. And it is OUR mistakes we need to fix, no one else's.

Good luck, and when you figure all this out, please help me.

Here is what Steve Lopez of Chessbase.com says

Play a lot. Serious, OTB chess. If possible, 80% of the time you spend on chess should be playing, and not so quick that you aren't really concentrating; I'd say at least 30 minutes / game. And after the game, go over it with your opponent, find out what you both were thinking and what sort of thing went wrong and why. Think of how you could avoid that type of mistake. Then play again.

The rest is for tactics exercises, endgames and playing through annotated games. And everything else you think is fun!

The point is that chess knowledge is just one part of what makes a good chess player; you need skills as well. Calculation, concentration, time management and so on are all best learned by doing and constant evaluation of what's going right and what's going wrong.

Thanks DrawMaster, ivoryknight and Scarblac!

Scarblac, your response really rings true to me.  I don't want to spin my wheels trying to improve, I want to feel (and see) that I'm really improving.  I will look to take your advice.  :-)  Thanks!

I have trouble learning from chess books because my attention span is sometimes limited in that genre.  I reached 2081 in standard on the chess server I play on, and it was because of video lectures.  *Nothing* has more successfully kept my attention and taught me so well as these gems.  Here is an article I wrote about my experiences with them:

http://www.associatedcontent.com/article/904029/chess_video_lectures_are_they_for_you.html?cat=19

Hope this helps!  :)

Thanks, Skwerly!

I have never watched a video lecture, but you have inspired me to seek them out.

An International Master (FIDE 2400) told me that he try to solve as much chess problems (mostly tactics and endgame) a day and try to analyse one game a day in detail. (I think the game analyses should include detailed analyses of your own games and those of masters playing the same opening as you).

Personally I dont found the computer analyses on chess.com useful so I rather use fritz to analyse it.

i haven't yet reach 1900 rating cause i'm new in this site but i've been playing chess since i was 10 or years old and i'm 48.   i used to play better players than me when i was that age and experience taught me that one has to play live chess, i mean playing chess in one sitting and joining tournaments.  i also read books, the classic ones because i believe one shoudl learn from the masters. 

1. Playing long OTB games
2. Analyzing your OTB games with your opponent
3. Analyzing your OTB games by yourself
4. Analyzing your OTB games with stronger players
5. Analyzing your OTB games with a computer
6. Solving tactical exercices (Tactics Trainer is good)
   
7. Going through annotated master games, trying to guess which moves have been played
   
8. Reading some good chess training book 
9. Learning important positions (Chess Mentor can help here)

And to add another point: play opponents that are 100-200 points above your own rating.

If you are +- 1900, the fastest way to become a 2000-2100 player is to play a lot of 2000-2100 players in serious games so that you get used to the level.

I have to agree with the draw master. You have to study your losses. I mean, I KNOW how incredibly difficult it is to do so, but you also must use a strong computer program to analyze. It is difficult to learn why to do the moves it suggests, but if you look at it for long enough and you have reasoning skills, then you should be able to figure it out. I hated studying my tournament losses, which is why i would take a few days, HOWEVER i DID put them in chessbase as soon as I got home form the tournament. So, study your losses, and I always found that if you play against people much better than yourself, you either A- Quit (But quitters never win) or B- Get better. So i guess it looks like you have a hard road ahead of you, but i believe with some shear determination you should be up there in no time.