This is about the comparison of terrestrial mammal species and marine mammal species, and how many subspecies each one has.

They have to model how the natural and artificial barriers in each environment produce subpopulations that are isolated enough to produce subspecies. And compare the simulations of the models with the data of real mammal species to see if they have similar results. This is not so close to pure math.

to follow up on gus_massa's point, it's not the "math" of an evolutionary thoery/agorithm that's being "proved" here; it's the assertion that it "proves" something about the environment it's trying to model

and saying you've "proved" something about natural events is probably bad practice

In those cases, is something about evolutionary science proven, or is a mathematical concept proven derived from a model of evolutionary systems? We can create proofs and do real math in extending and understanding the implications of the mathematical models we use for the world, but I don't think proving anything about the behavior of the model implies any proof of the behavior of the underlying science - it suggests it's likely but proof?