A useful concept from Brian Klaas’ Fluke: risk is when the outcome is unknown but the factors determining it are predictable, like rolling a die. Uncertainty is when neither the outcome nor the factors are knowable, say predicting a country’s GDP ten years out. Probabilities are useful in assessing risk, but in matters of uncertainty, they lend an air of false confidence.
Related: Frequency-type probability and belief-type probability. How often a coin comes up heads is a frequency-type probability; the probability is a measure of the frequency of a heads toss and converges on a more accurate measure with more tosses. Who will win an election is a belief-type probability: I can say I’m 74% sure that a newly discovered painting is an authentic Picasso, but it either is or it isn’t. The probability doesn’t reflect an event that can be repeated, and its accuracy is contingent on all of the assumptions I used in making it.



