The same thought crossed my mind. They're really not saying anything more profound than Richard Dawkins's "selfish gene" model, which has been around forever. Still the foundation of science is independent repeatability, and if they experimentally verify a "selfish gene" model playing out in real life, then, well, that's far from the worst expenditure of grant money I've ever heard of.
A computational model is far from real life - witness Dr. Hansen and his Global Warming charade. I suspect the researchers' model is not terribly sophisticated, which is probably why their conclusion is so banal.
Genes do not reproduce on their own - they require an organism and it is that organism which reproduces or not.
In order to test whether a gene is passed on, the entire set of genes and all their interactions and combinations within the organism must be accurately modeled. This in and of itself is a difficult problem. Additionally, there are a large number of such organisms competing (inter and intra species) within a given environment, which itself is constantly changing (naturally I might add, with another dirty glance at Dr. Hansen).
Thus, one must accurately model all the interactions and combinations of all the genes in all the organisms and then apply a fitness test against a changing environment.
It is a difficult thing to model accurately.