Quantifiable risk can be priced and sized for; true, unknowable uncertainty usually can't — and the two demand very different responses.
Economist Frank Knight's classic distinction, embraced in spirit by Munger and Buffett: risk describes situations where the odds, even if unfavorable, can be estimated with reasonable confidence — a coin flip, an insurance actuarial table. Uncertainty describes situations where the range of outcomes and their probabilities simply cannot be known in advance, no matter how much analysis is applied.
Every investment involves some uncertainty about what happens next — the question is how knowable that uncertainty actually is.
Ask whether the odds are measurable from history and stable patterns, or whether they are fundamentally unknowable right now.
Commit real capital to situations with measurable risk; avoid or drastically limit exposure to true uncertainty.
Decades of stable consumption patterns make future demand for candy and soft drinks a case of quantifiable, measurable risk — not the kind of unknowable uncertainty that would call for extreme caution.
Munger and Buffett have long avoided most early-stage biotech and speculative technology investments, judging the probability of any single outcome — success, failure, or disruption — to be genuinely unknowable rather than merely difficult.
Berkshire's purchase of the BNSF railroad was built on more than a century of stable, predictable infrastructure economics — freight volumes, route networks, regulatory patterns — a deliberate choice of measurable risk over speculative uncertainty.
Ask directly whether the key variables are measurable from history or fundamentally unknowable right now.
Favor industries and business models with long, stable operating histories over novel ones.
Treat "I can't put a probability on this" as a genuine warning sign, not a challenge to be overcome by more analysis.
Size positions much smaller — or avoid entirely — when true uncertainty cannot be reduced.
Remember that avoiding uncertainty is not the same as avoiding all risk; well-understood risk is investable.
We don't try to be smart about things we can't know — we try to be consistently not-stupid about the things we can actually measure.