How about if we count up all the known fossils in the world and divide that into the number of species known to have ever existed? We could come up with a ratio--the number of known fossils per species, on average.
It might be a chore counting up all the fossils, of course.
Also, there's the problem of identifying a new species. What one person might call a distinct species, somebody else might not. But I suppose one could ignore that.