[jamesrcole]

I think the problem is that people tend to focus on the size of the numeric probability, as if it was intrinsically meaningful. And don't think to consider contextual details like in the example you give. It's the combination of the probabilities to the contextual details they're applied to that really matter.

Another illustration of that is, if 99.99% equates to 1/10000 survivors, the meaning of it depends on how many bacteria exist on the area in the first place. It has quite a different meaning if the typical number is 10000 vs if it's 100000000