A question for those who know something not only about chess but about chess improvement:
Item 1: I have Fritz analyze all my recorded games.
Item 2: I’m kinda blunder-prone.
Do you think I’d benefit at all from making a file of positions in which I made terrible decision"s and compiling them into a “Choose the right move” book for myself? Or would this be an ineffective waste of time?
I am also trying to improve with the help of Fritz. Here are a few things that I have tried:
When the engine is running I have it display its top eight choices. If the move that I played doesn’t make the top eight then I’ll stop and look at the other moves that I should have considered. If the move that I played is Fritz’ fourth choice and 1/3 pawn worse, I don’t worry about it as long as I was playing with a plan.
When a late middle-game or endgame catches my interest, I’ll back up to a point where I think that I was clearly winning and play it out against Fritz a few times. I’d rather not admit how much of a lead I need in order to finish off the computer without seeing its analysis.
Recently I‘ve typed in my own analysis before seeing evaluations from Fritz. Often the critical turning points are not where I thought they were – something that I didn’t realize before trying this method.
Regarding the suggested blunder book, I don’t plan on trying that. One type of blunder is a missed tactic, and seeing tactics quickly might be more pattern recognition than it is calculation. Solving puzzles by tactical theme (knight forks, or pawn forks, etc.) would make you less susceptible to each tactic once it is recognized as a pattern. The few examples from your games are less likely to form a pattern. A puzzles book is more likely to cover more known patterns and variations of patterns than have happened by chance in your games. I enjoyed Neishtadt’s “Your Move”. I have not read many puzzle books so others may have better suggestions.
Sometimes the drop in evaluation is misleading: if Fritz goes from +3 to +1.6 but you now know how to win, then Fritz is “wrong.”
Here’s a technique that a 2300 learned from Kaidanov (OK, they didn’t talk about Fritz):
Take your last ten losses. Identify the losing move(s) with Fritz’s help (probably accompanied by at least an -0.5 drop in evaluation). Now you have at least ten positions (maybe more if you’re like me, saving lost games, then re-losing them). What common error do you see in these positions?
I did this a few years ago, and discovered three losses in which allowing a discovered check was my best (only) defense–in all three games, I had automatically rejected the best defense as too scary.
Here’s a training idea I have recommended to many beginning level players. Set the program for its strongest level of play and begin a game. Then, deliberately give away a whole bunch of pieces - Q, both Bs, maybe a N as well. Then use the command to switch sides so you’re playing with the extra pieces. Find the minimum material advantage you need to win every time against a super strong opponent. Play many games like this. The goal is gradually to reduce the material advantage you need to win every time.
– Hal Terrie
I tried it - and found that to win consistently I typically had to be up only one King. 
Well, if I knew something about chess improvement… I’d be improving. Or teaching. 
But seriously, it seems like a large amount of work for learning singular points in your play. If I were doing such a project, what would be critical to me is learning why making the right move was right in the circumstance and why other moves were wrong. i.e. go beyond move x is right in this position and simply choosing the right move and try to learn and apply why I’m choosing the wrong moves. (I think you already knew all this… 
)
One thing that I’ve been doing recently is playing a fair amount of Blitz training games against Fritz in Friend Mode. After most every game I copy the game over to a separate base, then do a bit of self analysis and then analysis with the engine. Once the wheel of my life slows down a little bit I’m going to step through as many of the games as I can to see if I can spot any patterns. (Besides that I lose quite a bit. )
This sounds kind of like what I had in mind. Perhaps I’ll try it.
This is a very good method: playing straight odds games against the computer (humbling as it is) works, too.