I'll add, most of my understanding was done this way: somewhat experimentally. Come up with a hypothesis of how things might work, figure out a way to prove it, and see if I was right.