"where I get to theorize an outcome and then build a model that will produce the outcome. Proof for me is evidence produced after the fact. "
OK. Svante Arrhenius kicked off global warming science. Using only pencil and paper, and putting months or years into the effort, he worked out an estimate of the climate effect of 560 ppm CO2 (double the 280 of his day. Currently, we're about halfway between the two.) His estimate has held up pretty well. We've seen a temperature gain about half what he predicted for 560.
Computer models are not moonshine. They're a compromise between pure physics, which would give exactly the right answer if only the numbers could be run, and budget reality, which does not allow unlimited number crunching.