Epistemic Humility
Epistemic humility is the discipline of matching confidence to evidence. In The Complete Collection - Howard Marks, it becomes especially visible during the COVID memo sequence: the future is important, but much of it is not knowable in advance.
Marks repeatedly separates:
- Facts - things currently known.
- Informed inferences - extrapolations from relevant analogies.
- Guesses - opinions presented with more confidence than the evidence deserves.
The practical lesson is not paralysis. Investors still have to act. But they should act with calibrated confidence, staged commitments, and awareness that big-picture forecasts are often least knowable precisely when they feel most urgent.
Why It Matters
Overconfidence is dangerous because people who lack the skill to judge a domain often also lack the skill to recognize their own limits. Marks applies this to pandemic forecasting, macro forecasting, inflation debates, and expert commentary.
Good judgment requires asking:
- Is this claim important?
- Is it knowable?
- Is the speaker an expert in this domain?
- Does the speaker express confidence proportional to evidence?
- What would change my mind?
Links
- decision-quality-vs-outcome - Acting well under uncertainty does not guarantee a good outcome.
- second-order-thinking - Requires seeing what the consensus assumes and where confidence may be misplaced.
- illusions-of-competence - Overconfidence often disguises shallow understanding.
Munger's Iron Prescription
Munger operationalizes epistemic humility through his iron-prescription: "I'm not entitled to have an opinion unless I can state the arguments against my position better than the people who are in opposition." He also distinguishes Planck knowledge from chauffeur knowledge — real calibration requires knowing which type you have in each domain. His circle-of-competence concept is epistemic humility applied to investing: know the boundary of what you understand, and don't act outside it.
Taleb's Randomness Discipline
Taleb adds a probabilistic version in fooled-by-randomness: humility means remembering that the realized path is not the whole truth. People over-explain outcomes, undercount alternative-histories, and mistake lucky survival for skill. Epistemic humility therefore requires inspecting sample selection, tail risk, and hidden skew before trusting any story.
Sources
- fooled-by-randomness - Luck disguised as skill, probability blindness, induction limits, and alternative histories.
- the-complete-collection-howard-marks - "Knowledge of the Future," "Uncertainty," "Uncertainty II," and "The Illusion of Knowledge."
- poor-charlies-almanack - Iron Prescription, Circle of Competence, Planck vs Chauffeur Knowledge.