Goes to show that actual possible futures have much wider uncertainty bounds than industry-standard projections like the ATB imply, and planning + analysis should reflect this (including more emphasis on wide bounding cases - something I've been trying to focus on recently)