It is unclear why sexually reproducing isogamous species frequently contain just two self-incompatible mating types. Deterministic theory suggests that since rare novel mating types experience a selective advantage (by virtue of their many potential partners), the number of mating types should consistently grow. However, in nature, species with thousands of mating types are exceedingly rare. Several competing theories for the predominance of species with two mating types exist, yet they lack an explanation for how many are possible and in which species to expect high numbers. Here, we present a theoretical null model that explains the distribution of mating type numbers using just three biological parameters: mutation rate, population size and the rate of sex. If the number of mating types results from a mutation–extinction balance, the rate of sexual reproduction plays a crucial role. If sex is facultative and rare (a very common combination in isogamous species), mating type diversity will remain low. In this rare sex regime, small fitness differences between the mating types lead to more frequent extinctions, further lowering mating type diversity. We also show that the empirical literature supports the role of drift and facultativeness of sex as a determinant of mating type dynamics.