It’s certainly a functional way to model many many systems in evolution, social sciences, economics, etc. But all models are limited at least by how much complexity you can put in.

The fact that humans actually don’t behave rationally itself is a huge discovery of game theory. Evolutionary models fundamentally rely on game theory in a way that is hard to overstate. Genes are inherently rational actors, the system is just complex.

I don’t doubt there are plenty of misapplications of it like anything else. But I mean, same with statistics, or calculus, or set theory.