What To Do When Nothing Has Happened
What To Do When Nothing Has Happened?
Raymond “Randy” Freeman
Process Safety Progress September 2011 (Vol. 30 No. 5)
How should you estimate the probability of some catastrophic event that hasn’t happened yet?
I could not find an open access PDF of this article, shipmates.
This is written from the perspective of a risk assessment, one that needs numerical values for probabilities. Freeman finds 13 different ways to assign a frequency (events per year for this paper) to some horrible thing that hasn’t happened.
The 13 ways to calculate a frequency all have some mathematical or statistical motivation. None of them are just guessing or arbitrary, which seems weird. Usually mathematical or statistical calculations are presented as “welp, here’s the only way we can see how it’s done”. This makes the choice of event frequency calculation a lot more like engineering and a lot less like science. You get to choose your heuristic.
Freeman advises using one of:
- λ = 1/(3n)
- λ = 1.3863/(2n)
- λ = 1 - 0.51/n
if you have 10 or more years (units of n) of data, a table in some obscure book if you don’t.