do we have interesting observation about statistics about human mistakes?

[Uri Blass]

For example possible observation(do not know if correct) may be the following:

Humans (with lichess blitz rating 2000-2200) do more mistakes when they do not push a pawn forward when the right move is to push a move forward relative to mistakes of pushing pawn forward when you do not need to do it.

Of course you can define something different than pushing pawns forward that I can call X when
observationstons should be always humans do more mistakes of not doing X(doing X) when they need to do the opposite relative to mistakes of doing X(not doing X) when they do need to do the opposite.

I wonder if these type of observations can help humans to play chess better.
Note that of course things may be different for different players but I would like to know generally about humans at specific rating range and if things change when you change the rating range.