lichess.org
Donate

How to Learn from Mistakes?

Most of the time, people respond to "What do I do if I lose?" with "The best thing you can do is to Learn From Your Mistakes" and, in a way, there is logic behind that. But, here on lichess, how are we supposed to do this? I'm here with Stockfish and he's showing all kinds of inaccuracies by myself and my opponent, and possible better moves in those positions, but the problem is I can't comprehend why the inaccuracies are bad and why the substitute moves are good, because Stockfish can't talk and explain why. We are humans and we need concepts, not computerized analysis and random moves with no explanation as to why they're played.

For example, here's a game I played in IM @Sparklehouse's Simul (I was black pieces): lichess.org/A7y71p7i
SF thinks 15...Be6 was an inaccuracy, but I'm just trying to hold onto my pawn. Instead, SF recommended 15...Be5, attacking white's h2 pawn. But after 16. Ndb5 Qe7 (again, more SF lines) black's idea of attacking h2 is gone. Also, why did black play Qe7? Why not Qd8 or Qd7, protecting d6? How are we supposed to distinguish the best move from a poor move? I'm just wondering how we amateurs are supposed to learn from our mistakes with a robot who can't verbalize. Please let me know what you think, I'd appreciate it.

Thanks,
FAJTV
15...Be6 is a mistake as it loses a pawn.

15...Be5 holds on to the pawn. 16...Qe7 is more active as it threatens Qb4 with attack on b2. 16...Qd8 or 16...Qd7 are passive.
@tpr Thank you for the clarifications on the moves, but still, I'm wondering how I can discover the reasoning on my own and not be dependent on stronger players if possible, or is it necessary?
These lines are difficult to see. You can find it on your own if you take your time, preferably on a real chessboard with tangible pieces. Some is pure logic, such as prefer active moves over passive moves.

By the way some of your problems with your weak pawn d5 stem from your plan 11...Nb6 and 12...d5. Better and more active is 11...Nc5.
The OP's question is an excellent one.
My take is he's asking can someone "really learn" simply by reviewing SF's suggested moves, without any input of "chess concepts" being verbalized.

A very broad question. Some will learn more than others. How the information is processed, how we "learn", resources available, there are so many factors. But it is a good question... what can be learned from SF's analysis?

Two points. SF only "see's" out to it's horizon. (The evaluation is always in flux, as the next move is made). It's +/- evaluation is always materialistic (assigning values to pieces/pawns in any given position).

One obvious learning tool is that it provides disclosure of mistakes and blunders that were missed in the game. For the most part, it is a question asked of "coaches" if that is a route of interest.

May as well add a 3rd point.
SF is not Zeus
Typically when one side has an isolated pawn, their compensation is through active piece play. From the looks of the game, it would seem that you had passive play and a great deal of your moves were reactive rather than proactive. You need to formulate a plan and when you react to an opponent's idea, generally counter-attacking is a good principle to follow, as seen with the move Be5 (the move itself might not be good in this case, but you should look for such moves). Perhaps there wasn't a particular grievous mistake in this game as you drew and the mistake is in choosing an opening or a position that gives you an isolated pawn. I for instance generally hate playing with isolated pawns and would much rather play against them. Identifying position types that you better understand or that have concepts that suit your style more is important.
As far as the evaluations are concerned when going over a game, I think computers are useful mainly in tactical positions. An example of this is move 16 for white, he should have played Nxe6 and it's justified with a tactical reason with Nxd5 being played after fxe6. However if you want to perhaps try to understand the moves stockfish spits out, you can follow the line that is given for at least a few moves and see if the comp has an actual plan. A lot of it is using basic precepts that are taught to be good in chess and using that knowledge to try to justify your moves. A simplistic example is rooks belong on open files, another one is that an active rook down a pawn is preferable to a passive rook with equal pawns on the board. You use your foundational knowledge to help understand why some moves/plans are good or bad. With comps, they can lead you down a meaningless path sometimes if they just spit out moves that have no connection to one another.
Great stuff by @NM Horobetz. He provides the necessary "verbalization."
The "tool" SF really is not of great value in the learning process, unless there exists a background in chess principles which SF does not give. I think new players will benefit more from books and CD's, where as experienced players are better equipped to use SF as a learning tool.
A generalization to be sure
The strongest chess programs are now beating humans at chess. We may have a good game comprehension, but the power of the machine is just too much for us.

For the good players, and by good I mean the best chess players in the world, their knowledge and experience is enough to play "almost" perfectly ( and that's why you see so many draws when you look at high rated player's games).

If we could play perfectly, every game of chess would end with a draw.

I often use Stockfish to analyse some tactical lines, or to study new openings. And I'm also often confused when the computer tells me to move my bishop on a square when i thought about putting it on an other one.

Using the computer as an advice is indeed a good way to understand and develop your "tactical sense" (I don't know how to call it), HOWEVER, you can't just rely on the machine.

Consider two strong machines playing against each other. Let them play, and look at the game. You should find crazy tactical lines, because the machines are playing "almost" perfectly, and end up in crazy positions, almost perfectly equal.

Studying the games Computer vs Computer may be interesting, but you should remember : you will always face a human opponent. Learning from the machine will tell you how to play vs the machine. From a human point of view, chess is slightly different. If I had to choose between two equal positions, which one would i choose ? I can't already know if one of the two will cause me some trouble 30 moves later if I miss a small line. I'll just choose the one where i FEEL better.

Learning from your mistakes is a matter of feeling. You can evaluate a position, but no magical voice will appear in your head and say "+0.2" (but rather "I don't feel like it. Remember that game ?").

I'm not a chess master (not yet, not yet), but i progressed A LOT in a short amount of time, trying to "feel my playstyle", and learning from what didn't work before.

Once you've got the "good feeling", you can work on calculations. Having both the "feeling" and the tactical force should make you progress for sure.

There's no magical formula. But no doubt your mistakes present-day will be the basics of your future success.

Have a nice day

This topic has been archived and can no longer be replied to.