Predicting a truly novel observation is rare, and has always been rare. Mostly what physicists do is clarify our understanding of things we've already observed,

Predicting a truly novel observation is rare, and has always been rare. Mostly what physicists do is clarify our understanding of things we've already observed, and extend techniques to handle cases that are more difficult to calculate than the cases that are presently considered tractable. For every one coulomb's law or socks-cling-to-stuff-after-drying effect there are a million papers about figuring out how to calculate the electrostatic fields of various configurations. The vast majority of what theoretical physicists do is "just math" except that unlike math, it's aimed at a problem posed by nature rather than a problem imagined up on the basis of what seems most interesting.