Study your defeats!

Your most valuable tool for improving at chess is your own lost games. This is the view of Mark Dvoretsky, one the world's best coaches. Analyse them in detail, preferably with  a chess computer or stronger player, and you will become a better player.

This may be true but I think is suffers from one serious flaw.

When we are studying or analyising we are not in the same situation as when we are playing. The two processes are different with study and analysis there is nothing at stake, no tension , no clock hence we are not in playing mode.

I find that turn-based is also tensionless chess, unlike live chess, hence I would suggest that that ones grade at turn-base should be 300 or even 400 higher than real OTB. This naturally will depend on the effort and time you put in to your moves I have time being retired but some on this site are short of time. Some indeed have far too  many games on the go to play at their best.

I save the PGN's from all my completed games here to study what worked and where I went wrong.  Usually I use Chessmaster to analyze the games or I'll use the free computer analysis here.  Sometimes it can be eye-opening to see how critical the computers can be of your moves.

I think a coach would disagree that you should use a computer to analyse the games. analyse them and find your own errors. After that use a computer as the last step. you are not a computer and do not think like one.
a computer can look at millions of positions and evaluate them.

I doubt that you think like a computer.

I think analyzing your games via an engine does NOTHING to improve your game...as a rule chess engines are much much stronger than you...you won't be able to determine WHY it chooses certain moves over others.preferably go over the games yourself or with someone slightly stronger than you...say 200-300 points.

it is easier to retain in ones memory a defeat for  avaluable lesson for improvement. dvoretsky endgame manual book  has many valuable lessons for improvement also, that we  can perfect our endgame skill.

You've obviously never used a recent version of Chessmaster, because it shows you which moves are better and why, playing out the moves how they'd likely be played, as well as explaining it in plain English.  It doesn't just tell you where you blundered, but shows you why.

I must say that I don't rely solely on computer analysis to try to improve my play, but it certainly doesn't hurt to see where you went wrong.

I've seen a lot of people say after they analyzed with the computer. On move 20 I should have played another move then I would be fine. the computer will tell better is...

I think it isn't that simple - your move 20 may have been part of a wrong plan and made sense to you while playing. So the wrong plan may have started on move 15.

move 20 may have overlooked some counter play the opponent could play. making your plan useless.

I try to work with my thought process that lead to a bad move instead of just pointing on one move.

I may have thought I could win material by playing this move and could not.

Obvious it is a flaw, but I think its a bit overkill to call it a serious flaw. I mean seriously, you get better at game type situations by playing the game and everything you do out of the game to get better on helps. 

It is very unlikely that any two chess games will be the same move for move unless the two players are following another game. So the positions you see in each game you play are unique you cannot plan to deal with the unique situation anymore than you can plan when driving to deal with an accident.

The vast bulk of analaysis simply makes the mind more fluid but games do that even better for they are real situations not pretend stuff.

Your mind will be limited by its speed and accuracy and you will have a limit under any given conditions. If you have a quick mind and a high intelligence and dedication you may rise to expert or even master.

Kasparov has an IQ of about 180 thats one in a million. The polgar sisters 170 and this is significant.