Full original question: Why is it so hard for humans to learn from their own mistakes considering A.I. imitates the human brain? Is Alphazero only better than humans because it played more games? Or is Alpha0 better at learning from its mistakes?
My thought process:
It obviously has played more games and it sure has an impact.
...
Considering that A.I. imitates the human brain it should repeat mistakes too (not only learn).
...
It only repeats mistakes at the beginning, but it does not once it identified something as a mistake.
...
A.I. only copies the good things from our brain. A.I. has no strange or bad habits, is not stubborn or anything similar.
...
If we want to be like an A.I. we need to reduce our bad habits holding us back from improvement.
...
New Question: How do we reduce our bad habits in chess?
...
Maybe if we reduce our bad habits in chess we will do in daily life too.
...
Example: I keep playing Ng5 thinking everytime that this will be a great attack. My opponent proves me wrong in most cases.
...
I cannot ban the move completely, since it might be good in some cases.
...
New Question: How do we indentify when our "habits" are actually good? By playing a lot of games and looking for when it actually worked out? Unfortunately not at my level, because if it works out it is more likely to be a sign that our opponent played bad and would have nothing to do with the position. The stockfish evaluation also doesn't really help sometimes, since having +0.3 advantage can mean nothing if it is only +0.3 in case you play perfectly for the next 15 moves.
...
We need a tool to know how successful a move was in chess history. We have that in form of databases and opening books. But those probabilites are only valuable for that specific position, not probabilities on how successful the move Ng5 is in general - having certain other pieces or pawn structures on the board.
Final question: Do we have something like that?
Also I would be very interested in your thought process
My thought process:
It obviously has played more games and it sure has an impact.
...
Considering that A.I. imitates the human brain it should repeat mistakes too (not only learn).
...
It only repeats mistakes at the beginning, but it does not once it identified something as a mistake.
...
A.I. only copies the good things from our brain. A.I. has no strange or bad habits, is not stubborn or anything similar.
...
If we want to be like an A.I. we need to reduce our bad habits holding us back from improvement.
...
New Question: How do we reduce our bad habits in chess?
...
Maybe if we reduce our bad habits in chess we will do in daily life too.
...
Example: I keep playing Ng5 thinking everytime that this will be a great attack. My opponent proves me wrong in most cases.
...
I cannot ban the move completely, since it might be good in some cases.
...
New Question: How do we indentify when our "habits" are actually good? By playing a lot of games and looking for when it actually worked out? Unfortunately not at my level, because if it works out it is more likely to be a sign that our opponent played bad and would have nothing to do with the position. The stockfish evaluation also doesn't really help sometimes, since having +0.3 advantage can mean nothing if it is only +0.3 in case you play perfectly for the next 15 moves.
...
We need a tool to know how successful a move was in chess history. We have that in form of databases and opening books. But those probabilites are only valuable for that specific position, not probabilities on how successful the move Ng5 is in general - having certain other pieces or pawn structures on the board.
Final question: Do we have something like that?
Also I would be very interested in your thought process