> Like lupatus mentioned in the other response, there is considerable variability in the risk from terrorism. "Considerable variability" doesn't begin to cut i

> Like lupatus mentioned in the other response, there is considerable variability in the risk from terrorism. "Considerable variability" doesn't begin to cut it. Deaths from terror attacks follow a power law. That's why it doesn't make sense to compare them to fundamentally linear phenomena like traffic or smoking deaths. Coincidentally (or not), flood levels also follow a power law. The thing about a power law phenomenon is that you can use it to justify an arbitrarily high prevention spend, because given the sample you have, you can guarantee an event lying outside it. Repeat ad infinitum.