Risk Assessment Scenarios

The formal distinction between risk and uncertainty was sharpened by economist Frank Knight in 1921. Risk applies when probabilities can be reasonably estimated (insurance actuarial tables, well-tested medical procedures); uncertainty applies when they cannot (novel technologies, unprecedented geopolitical events). Expected-value calculations work for risk but require care under uncertainty. Beyond Knight, the modern empirical literature on how humans actually perceive risk runs through Daniel Kahneman, Amos Tversky, and Paul Slovic, whose work showed that intuitive risk judgments are systematically distorted by availability (vivid events feel more probable), affect (emotionally loaded options bypass numerical assessment), framing (the same statistics produce different decisions when phrased as survival vs. mortality), and the certainty effect (moving from 99% to 100% feels disproportionately important).

Two specific calculation moves repay deliberate practice. First, cumulative probability: a 2% annual risk over a 50-year planning horizon compounds to roughly 64%, which transforms an apparently rare event into one that is more likely than not over the relevant timeframe. The mayor in this exercise who treats 2% as 'basically zero' is making a textbook error that engineers and actuaries are trained to avoid. Second, downside survival: when a bad outcome could end the game entirely — bankruptcy, ruin, mission failure — expected value alone is insufficient and you must additionally evaluate whether you can survive the worst case. Nassim Taleb's work on antifragility makes this point at length: pursuing the highest expected return without checking the survival constraint is a recipe for ruin in any sufficiently long game.

The precautionary principle deserves its own attention as a decision rule for a specific class of problems: catastrophic and irreversible downsides under genuine uncertainty about probability. It is not a general rule for all risk; applied indiscriminately, it would paralyze all innovation. But for irreversible interventions in complex systems (releasing self-propagating organisms, geoengineering at planetary scale, deploying autonomous systems with unbounded action spaces), the asymmetry between the bounded cost of waiting and the unbounded cost of an irreversible mistake reverses the default in favor of caution. As you work the scenarios, practice converting feelings about risk into numbers, calculating expected values and cumulative probabilities, and noticing when downside severity should override expected-value comparisons. For more on probabilistic reasoning, see the Probability & Statistics exercises.