Art. 03 · What a risk costs · 3 min read
Why the average is a fair basis for a decision
“It either happens or it doesn’t” is the commonest objection to risk maths. It is a fair objection, and it has a precise answer.
Flip a coin once. The average outcome, half a head, never happens. For a single flip, the average describes nothing you will actually experience, and anyone who tells you to plan around it is being silly.
Flip it five hundred times and something changes. The proportion of heads stops wandering and settles near half, and it does so reliably enough that you could take a bet on it. Nothing about any individual flip changed. What changed is how many of them there were.
Your business is not one flip. It is a few dozen risks, running across every month, for as long as you trade. Multiply properties, vehicles, employees and years together and you are well into the part of the graph where the average is a sound basis for planning.
But the objection is right in one specific case, and it is the important one. When a single event can end the business, you do not get a long run. You get one flip that matters. There, the average is actively misleading: an event with a 1-in-500 chance of costing you everything has a modest risk cost and an unacceptable outcome.
That distinction does most of the work in this series. Many small, independent risks: the average is decision-grade, and you should probably carry them yourself. One rare, ruinous risk: the average tells you almost nothing, and that is exactly what insurance is for.
Use it
- Sort your risks into two piles. "Annoying if it happens" and "over if it happens." They deserve completely different reasoning.
- Use averages on the first pile. Frequent, survivable losses average out. That is where risk cost is a reliable guide.
- Ignore averages on the second. For a business-ending risk the right question is not "what does it cost on average" but "could we survive it, and at what price can we hand it to someone else."