Police complaint after Paris weather sensor spikes fuel $35K bets

Météo France lodged a formal complaint with the Roissy Air Transport Gendarmerie Brigade after two brief temperature spikes at a sensor near Paris-Charles de Gaulle Airport coincided with well-timed wagers on the Polymarket platform.

The agency reported the anomalies occurred on April 6 and April 15. In each case the sensor recorded jumps of more than three degrees Celsius and returned to normal within minutes. Météo France cited physical checks of the instrument and analysis of logged data in its filing and classified the complaint as interference with an automated data processing system.

Multiple accounts placed large, low-probability bets on Polymarket markets that resolved based on the sensor readings. One account that placed what Météo France described as a stake of a few dozen dollars on April 6 received about $14,000 when the market settled in its favor. On April 15 an account spent about $120 on a low-probability position and cashed out roughly $21,000 within 30 minutes of the spike, a roughly 180-fold return.

A blockchain analytics platform that tracked the trades flagged the April 15 purchase as notable because it was placed minutes before the anomaly at odds below 1%. The analytics firm reported the account’s typical wager was about $6 at about 72% odds, while the winning trade was $120 at about 0.6% odds — a marked increase in bet size and opposite odds compared with the account’s history.

The April 6 winning account had been created two days before the bet, according to information reviewed by investigators. The timing of account creation and the pattern of wagers are part of the matters Météo France identified in its complaint.

Polymarket operates prediction markets that resolve when real-world data meet predefined conditions. The company has recently received a large investment from an established exchange operator and is pursuing additional funding. Regulators and lawmakers have recently considered restrictions on certain prediction market offerings, including sports-related markets.

A separate recent incident involving a live misread of a sporting outcome allowed a trader to convert a modest stake into a six-figure payout before the error was corrected. That episode and the sensor-linked trades have been cited in discussions about how market rules and data sources affect contract outcomes.

Mark Roulston, a researcher who studies prediction markets at Lancaster University, commented that contracts tied to a single reporting station can be vulnerable to faults or anomalies. He suggested that using averages across multiple stations or applying procedures to exclude outliers could reduce the chances that a single erroneous reading determines a contract outcome.

Météo France’s complaint focuses on whether the sensor was tampered with or otherwise interfered with during the periods in question. Investigators will examine the instrument inspections, the logged measurements and the trading records linked to the contested markets as part of the inquiry.