Sharpen your judgement under uncertainty
Train your reasoning, not just your recall. Most tools reward knowing the right answer. Judgement Gym develops knowing how confident to be in your answer — and being able to explain why. An AI-powered coach that stress-tests your thinking.
AI Tutor
A direct, curious sparring partner that probes your reasoning with base rates, crux identification, and steelman challenges — only revealing the answer at the end.
Calibrated Scoring
Separate reasoning quality from outcome accuracy. Brier scores measure calibration while LLM evaluation assesses your actual thinking process.
Track Progress
Monitor your confidence calibration, reasoning quality trends, and revision patterns across empirical, normative, and strategic scenarios.
How a rep works
Face a scenario
You're presented with a real-world case — neutral framing, no hints. Form your own position first.
Commit your position
Lock in a falsifiable claim and confidence percentage before the tutor pushes back.
Get stress-tested
Three sequential probes challenge your reasoning: base rates, crux identification, and steelmanning the opposite view.
Debrief and score
See how your reasoning held up. Review your calibration, the moves you used, and the ones you missed.