I meant that tests per million is a good enough measure of randomness to justify saying that the positivity rates in countries with different positivity and tests per million are probably not comparable. “Fermi math” refers to approximating from little data, something for which Enrico Fermi was known. (Why the method deserves a name isn’t clear to me, but it has one nevertheless. I think “back-of-the-envelope” captures the idea, as does Wikipedia.)