Thinking about careers that use statistics? The job title “catastrophe modeling assistant” caught my eye recently in a job announcement. It turns out there is a whole community of catastrophe modelers, or “cat modelers” in the jargon. The insurance industry relies on probabilistic modeling of catastrophic events to price insurance, and typically turns to expert consultants for these models. AIR Worldwide and RMS are two major sources of catastrophe models.

Thinking about it probabilistically, the advantages of such a line of work are evident.

If you are forecasting the weather, it will not take too long to determine whether your “30% chance of rain today” models are in the right ballpark. 50-100 such forecasts should do the trick, and a year or so should suffice to accumulate them.

If you are forecasting the probability of a catastrophic typhoon hitting Manila, though, there’s no way you will accumulate sufficient evidence to prove or disprove the model during the lifetime of the modeler, let alone the model.