In the last weeks of the 2024 US election, when nearly every public poll was calling the race a coin flip, a single anonymous trader in France was quietly betting that the whole industry was wrong. He was not a pundit and he did not go on television. He just kept buying, and buying, until his wagers alone were bending the odds on the largest prediction market in the world.
He is known only as Theo, a former trader who asked to stay anonymous, citing his family and his safety. What made him remarkable was not the size of his bankroll but the reason he was so sure. Theo had become convinced that the polls were missing something human: that a slice of Trump voters simply would not tell a stranger with a clipboard the truth. So he paid for his own research, built around a question that sounds almost too simple to matter.
The question that beat the pollsters
Instead of asking people who they would vote for, Theo's private pollsters asked who they thought their neighbors would vote for. The technique, sometimes called neighbor polling, is designed to strip out the embarrassment factor, the quiet voter who guards their own answer but happily guesses about the house next door. Across those private polls, the neighbors kept leaning harder toward Trump than the headline numbers did. Theo decided that gap was real money, and he pressed it.
He started with a stake of around thirty million dollars and, after his late-October data firmed up, pushed his exposure toward eighty million across roughly a dozen crypto wallets. Those bets were large enough to move the board on their own, which drew accusations that he was manipulating the market's read on the race. Theo waved the politics away and said he was in it purely for the profit. Polymarket, for its part, credited the neighbor-poll method with surfacing a signal the professionals had filtered out.
What the win really cost the industry
When Trump won, the blockchain-analytics firm Chainalysis put Theo's net winnings at close to seventy-nine million dollars. The number is staggering, but the more interesting damage was reputational, and it landed on the pollsters, not on him. A lone amateur with a contrarian survey had out-read an entire profession, and he had done it in public, one on-chain trade at a time.
The aftermath was less glamorous than the payout. French regulators, unamused that one of their nationals had turned an election into a personal windfall, moved to squeeze Polymarket out of the country entirely. Theo, characteristically, stayed out of sight.
The lesson traders took from him was not "bet big." It was that the crowd's confidence is only as good as the question underneath it. Theo did not have better nerve than the market. He had a better question, and on a prediction market a better question is the only edge that pays.
