When people do this kind of predictions, they often driven by emotional reaction. Best thing to switch actual evaluation on certain hypothesis is to make actual risks cost something.