Risk is not the same as uncertainty

A forecast that promises "73 percent chance of recession" sounds rigorous, scientific, reliable. Often it is precisely the opposite of rigour: genuine ignorance, dressed up with a number. There is a distinction over a century old that protects you against this, and once you have it, you read every forecast differently.

Two kinds of not-knowing

The economist Frank Knight drew the dividing line in 1921. There is risk, and there is uncertainty. Risk is measurable: you know the possible outcomes and their probabilities, as with a die, a roulette wheel or a mortality table. Insurers live on risk. They know the odds precisely enough. Uncertainty is something else entirely: the outcomes, or their probabilities, are genuinely unknowable. Will there be war? A technological breakthrough? A crisis, and when? No reliable number can be pinned on that. And the uncomfortable part is this: almost everything that truly matters about the future is not risk but uncertainty.

The false precision

Here lies the great mistake, and you see it everywhere: treating uncertainty as if it were risk, putting decimal places on the unknowable. "73 percent chance of collapse", the timeline of a new technology pinned to the year, the models that assumed the housing market could not fall. The number borrows an authority it has not earned. John Maynard Keynes put it bluntly in 1937: about the questions that matter most, "we simply do not know". The honest move, then, is not to feign precision, but to admit that. And then to act anyway, on the strength of judgement, not pseudo-mathematics.

How to read a forecast

That gives you a practical test. When you come across a forecast, ask: is this a calculable risk, a real base rate, repeatable events, or is this uncertainty in a suit? The answer tells you how much weight you may hang on it. A weather forecast for tomorrow sits close to risk. A forecast about where the economy or politics will stand in five years is largely uncertainty, however confident the speaker may sound. And as the section on rise and fall showed: dressing uncertainty up as risk is precisely the doom-monger's trick.

Take the difference in practice. An insurer pricing car damage leans on millions of cases that recur every year, that is risk, and his percentage means something. A commentator pricing "the chance of the next financial crisis" leans on a handful of events that never repeat in the same way, that is uncertainty, and his percentage means almost nothing, however firm it sounds. The same kind of number, an entirely different ground beneath it.

The other side

And to be fair: the opposite error is just as real. "We can know nothing" soon becomes an excuse for paralysis, or else for recklessness. "It is unknowable anyway, so I'll just guess". Some things are estimable, and refusing all estimation is its own kind of folly. The art is calibration: knowing the difference, and not letting one dissolve into the other. Pascal helps here: if you do not know the odds, weigh the stakes, what happens if you are wrong? Sometimes an improbable disaster with an irreversible consequence deserves more attention than a probable one with a small one.

The goal, then, is not to stop estimating. It is to estimate what is estimable, to admit honestly what is not, and to distrust most of all whatever hides its uncertainty behind precision.

Something to take with you: the next confident forecast you hear, about the market, politics, technology, ask yourself: is this a calculated risk, or uncertainty in the clothes of a number? And how much of your own plans rests on a precision that is not there?