April 8, 2026
Base Rate Neglect: The Calibration Error Almost Everyone Makes
A test is 95% accurate. You test positive. What are the chances you actually have the disease?
Most people say 95%. But if the disease affects 1 in 1,000 people, the real answer is about 2%.
The mistake is called base rate neglect: ignoring how often something happens in general when evaluating a specific case. Kahneman and Tversky first documented it in 1973, and it remains the most common calibration error in forecasting research.
Why it happens
Your brain substitutes a hard question ("what is the probability, given the base rate and the evidence?") with an easier one ("does this feel likely?"). The answer to the easy question feels like the answer to the real question. It is not.
This shows up constantly in professional decisions. Investors evaluate pitches based on how compelling the founder sounds, ignoring that ~90% of seed-stage companies fail. Hiring managers rate candidates on interview performance, ignoring that unstructured interviews predict job success with only ~14% accuracy.
The fix: anchor, adjust, record
Philip Tetlock's research on superforecasters found that the best predictors consistently start with base rates before adjusting for specifics.
1. Anchor. Before asking "will this work?", ask "how often do things like this work?" Find the historical frequency. That is your starting point.
2. Adjust. Layer in what makes your situation different. Superforecasters typically adjust 5-15 percentage points from the base rate, not 50.
3. Record. Write down your estimate, the base rate, and your reasoning. Without a record, you cannot calibrate over time.
Of 1,900+ people who took our calibration quiz, 44% scored overconfident. Base rate neglect is the most common contributing factor. If you want to see where you stand, the quiz takes about two minutes.
Related reading
- The Overconfidence Problem: Why Your 80% Sure Calls Are Wrong 40% of the Time
- What Is a Brier Score and Why It Matters for Your Decisions
- The Pre-Mortem: How to Find Fatal Flaws Before You Commit
Sources: Kahneman & Tversky (1973), "On the Psychology of Prediction." Tetlock (2015), Superforecasting, Ch. 7.