The notion of naturalness is important in particle physics, especially supersymmetry. I was a little surprised, then, that the idea, or at least the name, is apparently only ~30 years old (t'Hooft, 1980).
I know that Bayesian statistics, which formalises naturalness arguments with Bayesian evidence, is ~300 years old, and that with it one can elucidate connections between naturalness and Occam's razor (and even falsifiability). I don't know when those insights were first made, but surely not after t'Hooft?
Are there applications of the naturalness principle/fine-tuning in physics (or science) that significantly predate this t'Hooft definition? or even Bayes? I should add that I consider naturalness to be distinct from Occam's razor.