Computing & AI · the experiment
Be the clear-eyed onlooker
Strangers betting on other people's goals, prices to check against what really happened. Four questions, from beginner to expert, on what the crowd misjudges.
1 · Compare price and reality
Beginner
In 6,150 "Will I…?" markets, how did final prices compare with what happened?
2 · Count the missing YES
Recalibrating final prices: a self-market priced at 50% resolved YES 39% of the time; an external market at the same price, 47%. Across all self-markets and all prices, the average gap is 4.4 points.
Intermediate
Out of 400 self-markets each priced at 50% at close, how many fewer YES than the prices promise?
3 · Price the named misfortunes
294 self-markets have, as their YES, an outcome the person would not want: relapsing, breaking a streak, falling ill.
Expert
How were these markets priced at close?
4 · Judge the formula fix
The authors learn a correction from markets created and resolved before 2025, then apply it to later self-markets. The yardstick: the Brier score, the mean squared gap between price and outcome (lower is better).
Expert
What does the correction achieve, and why?
Crowd recalibrated
—