The concept of risk has entered into physics courses in various guises. It is treated explicitly in the "Advancing Physics" [1] course and implicitly at GCSE through "Ideas and Evidence" discussions. This could easily lead to such ideas as the balance between treatment and risk in radiotherapy and the likelihood of an asteroid strike. In this article we address a special form of statistics that is necessary when dealing with rare events. Whilst the mathematical details are unlikely to be necessary in the classroom it is hoped that this will provide background information and extension material relevant to this sort of question, which is all too likely to come up in the classroom. (Contains 2 figures and 1 footnote.)