Superforecasting
The Art and Science of Prediction
The big idea
Predicting the future is a real skill you can measure and train — and the people who do it best aren't the loudest experts. They're ordinary, humble types who think in probabilities and keep nudging their estimates as new facts come in, instead of committing to a story and defending it.
Why it matters
Tetlock's earlier research was brutal: he tracked pundits' political predictions for two decades and found the average expert was barely better than chance — his line was that they did about as well as a dart-throwing chimpanzee, and the more famous the expert, the worse they tended to do. Then came the good news. In a US-government forecasting tournament, Tetlock's Good Judgment Project fielded teams of volunteers — a retired computer programmer, a pharmacist, a homemaker — and the best of them, the 'superforecasters,' beat everyone else, reportedly even outperforming intelligence analysts who had access to classified data. The edge wasn't IQ or secrets. It was method: they broke a vague question into smaller answerable ones, started from base rates, and updated in small increments rather than lurching. Tetlock borrows Isaiah Berlin's fox-and-hedgehog split — hedgehogs know one big thing and force everything through it; foxes know many small things and stay flexible. Foxes forecast better.
Use it today
Next time you hear yourself say something 'will definitely' or 'will never' happen, put an actual number on it — 70%, not 'probably' — and jot down what would make you change that number. Then when news arrives, move the number a little instead of defending the call. A prediction with a percentage and an update rule is a forecast; without them it's just a mood.
Drop it in conversation
“There was a government forecasting tournament where a handful of amateurs beat professional intelligence analysts who had classified files. The difference wasn't brains or access — they just thought in probabilities and kept updating, instead of picking a side and digging in.”
In real lifeA project manager who says 'about 70% we ship by Friday' and revises it as blockers clear gives their team something useful to plan around, while the one who insists 'we'll definitely make it' right up until it slips leaves everyone blindsided. A weather app that says '30% chance of rain' is being honest in a way that 'it won't rain' never is — and you pack an umbrella accordingly. Someone forecasting a sales quarter does better starting from 'what usually happens at this point' than from a gut feeling about this one being special. The habit is the same across all of them: a number, a reason, and a willingness to move it.