But you can reproduce the probability distribution aspect of it quite reliably. If you think the dual slit experiment presents a problem to the scientific method or empiricism you're quite wrong.