Paul Krugman writes that you can't think correctly about climate change unless you take adequate consideration of relatively unlikely scenarios for disaster. Psychologically, if there's only a one or two percent change of something happening, we tend to put it in the "not going to happen" file. But sound policymaking would consider a one percent chance of a scenario in which billions die to be something worth worrying about.
I've been thinking about this lately as I re-read Watchmen inspired by the trailer for the forthcoming film adaptation. It's a reminder that in the 1980s, and especially before the Reykjavik Summit there was enormous anxiety about nuclear war. In particular, the criticism that Ronald Reagan's policies were likely to lead to a nuclear war was, though clearly not embraced by a majority, a fairly widespread and mainstream opinion. This comes out in Watchmen and The Dark Knight Returns and "99 Red Balloons" among other places:
Obviously, in retrospect this was wrong. Consequently, from a certain point of view the Nuclear Freeze Movement people and critics of Reagan's "evil empire" rhetoric look foolish. But when you think in terms of probabilities this isn't necessarily right. Say that under Jimmy Carter's policies there would have been a one percent chance of a nuclear war in the 1981-84 period whereas under Reagan's more aggressive policies there was a three percent chance of such a war. That still gives a very good chance that Reaganism will work out -- 97 percent is good odds -- but still probably means that the Reagan option is a bad idea. But then of course things will probably turn out okay, making the skeptics look foolish, and perhaps unduly biasing future policymaking toward aggressive options.