Probability estimates with zero Occurrences

On page 100 an error has been made in the cruise ship example used to illustrate the application of Quigley and Revie’s theory for the estimation of probability when there has been no prior experience of Occurrences.

The text from and including the sentences beginning “The Exposure, n is …” to “It is interesting to note…” should be replaced by this:

The number of trials, n is 30 x 100 = 3000 [cruises]

The probability of a damage consequence occurring on a cruise is therefore estimated at 

p = 1/(2.5 x 3000) = 1.33 x 10-4 [damage cases per cruise]

and, with an Exposure, e, of 100 [cruises per year] the Frequency, f, is estimated at

f=p x e = 1.333 x 10-4  x 100 = 1.333 x 10-2 [damage cases per year]

It is interesting to note that this is approximately half my very subjective estimate in Table 6.4.

Leave a Comment