HIP-4 Research

When the crowd was sure, and wrong

A calibration curve says whether prices come true on average. This is the other half: the markets where they did not. Ranked by what the side that WON paid on average: the lower that price, the more certain the crowd was of the opposite.

Cheapest winnerThe volume-weighted average of every buy on the side that ended up winning. 24¢ means buyers of the eventual winner paid 24 cents on the dollar, on average, across the market's whole life.

12.7¢

average paid on the winning side of NFL: Los Angeles Rams vs Arizona Cardinals

Largest by volume

$7.07M

2026 World Cup Champion: Spain, winning side at 37.3¢

What the crowd paidThe same average for the side that lost. The two do not add to 100¢: they are averages over different trades at different moments, not one book at one instant.

85.9¢

average on the losing side, 5,399 trades

The cheapest winning sides in the archiveEvery settled market with at least $250k traded and 50 buys on the winning side.

MarketWinner paidLoser paidTraded

A market is here because its winning side was cheap, not because someone called it: the buyers at 24¢ were a minority in a market that spent its life disagreeing with them.

Method, and what it does not say

Every buy of a settled market is placed by the price it paid and the side it bought. The side that won is averaged across its whole life, volume-weighted, not the closing price.
Ranked on the WINNER's average price, never the loser's. A market where one side is near-certain has a heavily traded cheap loser and nothing surprising in it.
Floors: at least $250,000 traded and 50 buys on the winning side. A thin market can print any price and is not a crowd being wrong.
This is not a claim that anybody predicted these. An upset is a market whose price was far from its result; whether the buyers at 24¢ knew something or were simply many, the archive cannot say.