The music I make is experimental. Which is to say, I don't know how to play, and I'm figuring it out by trial and error.
Often I get a spectacular result, only to find out it's irreproducible. Out it goes.
There's precious little theory involved in any of it, up front. I don't like to devise theories and then try to write songs that validate them. I prefer to wait until I've got some hard results, then see what theories can fit.
Bad musicianship, maybe. But good science.