Prediction Markets & the Crowd
Pool many independent guesses and the errors cancel — which is why a betting market’s price is often a sharper forecast than any expert.
You want the best possible forecast of some uncertain event — an election, a product launch, a scientific result. Your instinct is to find the single smartest expert and ask them. But there’s strong evidence that a crowd of ordinary, independent guesses — especially when people have money on the line — routinely beats the lone expert. That sounds absurd: how can a mob out-forecast a specialist? Understanding why (and when it fails) is the capstone of forecasting method. Hold the question: how could averaging a bunch of so-so guesses produce a great one?
Independent errors cancel out
The wisdom of crowds: when many people make independent guesses, their errors point in random directions — some too high, some too low — and largely cancel out when you average them, leaving something close to the truth. The classic jellybean-jar experiment shows the average guess beating almost every individual, including the confident ones. It’s the same idea as calibration meeting statistics: no single guess is reliable, but the aggregate squeezes out the noise (lesson 2) and keeps the signal. The crowd isn’t smart; the averaging is.
A prediction market turns the crowd into a price
A prediction market supercharges this: people bet on outcomes, and the market price becomes a live probability. If a contract that pays $1 if an event happens trades at 70¢, the crowd is effectively saying “~70% likely.” Two forces make it sharp: money means people only bet when they think the price is wrong (so information gets pulled in), and anyone who knows something can profit by correcting the price. The result is a constantly-updating, often well-calibrated (lesson 25) forecast — the price literally is the crowd’s probability, and it moves the instant new information appears.
Reading a market like a forecast: • A market contract on “Product X ships this year” trades at 30¢. • Translation: the crowd, with money at stake, puts it around 30% likely. • News breaks that a key part is delayed; traders sell, and the price slides to 12¢ within minutes. • You just watched a probability update in real time — faster and often better-calibrated than a pundit issuing a fresh op-ed.
When the crowd fails: the independence condition
The magic has a precise failure mode: it needs guesses to be independent. The moment everyone copies the same source, follows the same herd, or panics together, the errors stop cancelling — they stack, and the crowd becomes a mob (think bubbles, from the finance track). Correlated errors are how crowds get things spectacularly wrong. Thin markets (few participants) and manipulation can also distort the price. So the rule is: aggregation is powerful when guesses are diverse and independent, and dangerous when everyone’s thinking the same thing. That single condition tells you when to trust the crowd and when to be the one who thinks for themselves.
Picture everyone at a fair guessing a cow’s weight. Individually they’re hilariously off — but their guesses miss in every direction, so the high and low errors cancel and the average nails it almost exactly. Now imagine instead that one loud “expert” shouts a number first and everyone just echoes it: the guesses are no longer independent, the errors all lean the same way, and the average is as wrong as the expert. Same crowd, opposite result — the whole trick was that the guesses didn’t copy each other. A prediction market is that cow-weighing crowd, but with money making everyone guess their honest best.
Trusting vs distrusting a crowd forecast: 1. A deep, active prediction market with thousands of independent bettors puts an event at 65%. Diverse, independent, money at stake → treat it as a strong forecast. 2. A tiny forum where everyone read the same viral thread all “predicts” the same outcome → errors are correlated, not independent → discount it heavily. 3. Same act (aggregating guesses), opposite trust, decided by one question: are the guesses independent? When they are, lean on the crowd; when they’re all echoing, be the independent thinker.
This is the reading. The interactive version — active-recall quiz, a hands-on experiment you run in your own AI, and an earned mastery check — is free in the app.
Start this lesson free →