I saw this insightful tweet by the (possibly Singaporean) quant SystematicLS the other day:
One thing I would like to point out here that I think is a salient point is that having theory reduces the search space. That is an inescapable fact of finding models or signals.
— systematic longshort (@systematicls) August 12, 2024
One could traverse the search space quickly, test a lot of models, expect that a supermajority are… https://t.co/UYXtCiXq8P
… noise, and deal with that. Or one could use theory to bound the search space, and when a model is found, it’s less likely to be a false positive. True for discoveries as it is for signals and models.
I think this generalizes to a neat but obvious idea of math in programming. The search space of computation is infinite and unbounded. Math or theory in this case cuts a swift path in the space and forms a thread between a meaningful starting point and an end result.