"1 occurence" is misleading, as the miles driven is the actual random event. If you throw a coin 100 million times and 1 time it lands on its side - then you can have a rather reasonable estimate on how rare that possibility is even though you have just 1 occurrence. I.e., it may be that the chance is 0.5 per 100m or 2 per 100m; but it'd be very strange if the actual chance is 10 per 100m.