Someone should really create a dataset comparing predicted damage/casualties of storms vs. actual outcomes. It would really help set context for stuff like the panic around Milton, stoked by entertainers masquerading as meteorologists (such as this gentleman: [1]).