How to Think About Risk
Separate the probability of an outcome from the size of its consequences — and respect ruin.
Also known as: Reasoning about risk
intermediate Synthesis of decision theory and behavioural economics
Thinking clearly about risk means holding two things apart: how likely an outcome is, and how bad (or good) it is if it happens. Most risk mistakes come from collapsing those two, or from ignoring the special category of outcomes you can't recover from.
What it is
A useful first move is to split every risk into probability and consequence. A likely-but-trivial outcome and a rare-but-catastrophic one are completely different problems, even if a naive "risk score" treats them the same. Expected-value reasoning multiplies the two, which is right for repeated, survivable bets — but it breaks down at the extremes.
The extreme that matters most is ruin. Some losses are unrecoverable: bankruptcy, death, a destroyed reputation, an extinct species. For these, no upside justifies the bet, because if you don't survive the downside you don't get to keep playing. This is why a margin of safety exists and why you avoid strategies that risk catastrophe for incremental gain, however good the odds look.
Two biases distort risk badly. The availability heuristic makes vivid, well-reported dangers feel more likely than they are, while quiet statistical risks feel safe. Loss aversion makes us over-fear ordinary, survivable losses and shy away from favourable bets. Good risk thinking corrects both: reach for base rates instead of dramatic examples, and judge survivable decisions by their expected value rather than the sting of a possible loss — while treating ruin as a line you simply don't cross.
Worked example
Two founders each get a chance to double their company's value on a coin flip — but the losing side means bankruptcy. The expected value is positive, yet taking it is a mistake: a 50% chance of ruin ends the game, and you can't average your way back from zero. Contrast a marketer running many small, independent A/B tests where each loss is trivial — there, playing the positive-expected-value odds repeatedly is exactly right. Same maths, opposite decision, because one risks ruin and the other doesn't.
Failure mode — when it misleads
The framing misleads if you treat all risk as ruin (becoming paralysed and never taking survivable, favourable bets) or treat no risk as ruin (blindly maximising expected value into a catastrophe you can't recover from). The skill is telling the two apart. It also misleads if you estimate probabilities from vivid anecdotes rather than base rates — a well-calibrated sense of the odds is doing half the work.
How to apply it
- Split the risk: how likely is it, and how bad is it if it happens?
- Ask first: could this outcome ruin me? If yes, no upside justifies it.
- For survivable, repeatable bets, reason by expected value.
- Check the odds against base rates, not the scariest example you can recall.
- Keep a margin of safety sized to the uncertainty and the stakes.
Related entries
Composed of
- Margin of Safety Mental model Build in a buffer so that being wrong — and you will be — doesn't ruin you.
- Availability Heuristic Cognitive bias We judge how likely something is by how easily examples come to mind.
- Loss Aversion Cognitive bias A loss hurts about twice as much as an equivalent gain feels good.
Sources & further reading
The Black Swan
by Nassim Nicholas Taleb · book
Taleb's work on rare, high-consequence events underpins the emphasis on ruin.
Affiliate link — we may earn a commission, at no extra cost to you.