It depends on your definition of "confident". In the case of those people I used for my samplings I am "confident" of their types, since I observed them for a long time, which doesn't mean "100% certain". Of course some of those typings may be wrong. I think that most of them are right, but there's no way to prove it, which is the problem we keep coming back to.

At least everyone seems to agree that ESFps are relatively common.