Both are applicable. One cannot test unobserved, unrecorded phenomena but can only make inferences from what exists. One can also dismiss all phenomena as "occuring naturally" whether it occurs naturally or not. In either case, the assertions are no more scientific, and no more provable, than those of ID proponents.
The phenomena in question are not "unrecorded" as I pointed out to you. And they do occur naturally as they are occuring even to this day. You're dancing around the issue.