But this is in deployment software, i.e. I interpret your assertion as meaning "bug prevalence in software after it has been debugged is the same regardless of type system." My argument is that, at least for me, code becomes bug-free faster with static types than with dynamic, not that it ends up more bug-free in the end.