Debiasing

A deliberate practice for correcting predictable errors of judgment: slow down to engage reflective thinking, start from the outside view, reason in probabilities, seek the case against your own conclusion, update as evidence arrives, and keep score so the practice can improve. It does not try to make intuition unbiased; it surrounds intuition with procedures that catch its systematic mistakes.

Composed Of

• System 1 and System 2 — the reason the practice is needed: fast automatic judgment is where the biases live, and deliberate thinking is what corrects them.
• Overconfidence — the bias the whole practice is calibrated against: certainty that outruns accuracy.
• Planning Fallacy — a flagship error the outside view exists to counter.
• Reference-Class Forecasting — anchor on a base rate of similar cases before adjusting for specifics.
• Fermi Estimation — break an intractable question into estimable parts with visible assumptions.
• Probabilistic Thinking — state beliefs as granular odds that can be weighed and checked.
• Active Open-Mindedness — treat beliefs as hypotheses and look hardest for the disconfirming case.
• Bayesian Updating — revise in proportion to the evidence, neither anchoring nor overreacting.
• Calibration — keep score so confidence can be checked against outcomes and the practice can improve.

Why it Matters

Two sources reach the same conclusion from different ends. Thinking, Fast and Slow shows that bias is systematic and largely immune to willpower or awareness alone, so individuals cannot simply decide to be rational. Superforecasting shows that accuracy nonetheless improves when people adopt specific procedures: outside view first, granular probabilities, deliberate updating, and feedback. The lever is not a better brain but a better process wrapped around the same brain.

Signals

• Decisions made on the first confident story, with no outside view or base rate.
• Forecasts and plans that are never scored, so the same errors recur unchecked.
• Disagreement settled by who is most certain rather than who is best calibrated.

Risks

Procedure as theater, where checklists are followed but conclusions are not actually revisited; analysis that never converges because open-mindedness has no stopping rule; debiasing one decision while the surrounding incentives still reward confident, fast answers.