Show HN: The Forecaster Test- based on Tetlock's superforecasting research

1 point

4 months ago

In Tetlock's 20-year Expert Political Judgment study, domain experts performed no better in their own field than outside it — and "hedgehog" specialists actually did worse when predicting within their specialty. The Good Judgment Project (Tetlock & Mellers) later found that superforecasters weren't dramatically smarter than average participants — what separated them was thinking style. What predicts accuracy? Calibration, Bayesian reasoning, and willingness to update beliefs. These traits beat raw intelligence and domain knowledge. I built a 10-minute test that measures them:

Bayesian updating: Commit to a probability, receive new evidence, update. Tests whether you move the right direction and magnitude. Calibration: When you say 80% confident, are you right 80% of the time? Replication prediction: Predict whether famous psychology studies replicated. Instant feedback since outcomes are known.

Free, no signup. Results show where you're calibrated vs. overconfident, and how you handle disconfirming evidence. Looking for feedback on questions and scoring. Research basis is Tetlock, Mellers, Baron, and Stanovich — but I'm sure there are improvements to make.