A seven-months-pregnant writer in Seine-Saint-Denis spent last week trying to outrun the heat in the poorest département of mainland France, and she lost. Paris had just come through what was widely reported as the most extreme heat in decades, part of a European heatwave that produced France's hottest day since national records began in 1947. Paris itself registered a June record of 40.9°C. The 2003 European heatwave, by comparison, killed nearly 15,000 people in France. And while families across the city were sleeping on tile floors and rationing cold compresses, a small contract on Polymarket was quietly pricing the peak temperature for 30 June at a single weather station on the northeastern edge of town.

That contract is narrow, technical, and unusually instructive. It is not a market on heatwave mortality, or on climate policy, or on whether the French state is doing enough. It is a ladder of temperature bands, each one a separate yes/no bet, all resolving off one number: the highest reading on 30 June 2026 at Paris-Le Bourget Airport, as logged by Wunderground.

What the ladder is actually pricing

The highest-temperature-in-Paris market on Polymarket is built as a set of one-degree Celsius bands stretching from "24°C or below" up to "34°C or higher". As of the 30 June snapshot, the 28°C band sits at 41%, the 29°C band at 33%, the 27°C band at 22%, and the 30°C band at 7%. Everything else trails into the low single digits or below one percent. The market has a clear favourite, and that favourite is a reading right in the upper-twenties.

That is worth flagging for two reasons. First, the band concentration. More than nine in ten cents of probability are stacked across just three adjacent rungs, 27 to 29°C. Traders are not arguing about whether the day will be hot. They are arguing about which specific degree the peak lands on. Second, the 24-hour movement. The 28°C band climbed eleven points in a day while the 29°C band shed about five, suggesting that whatever forecast feed traders are watching nudged the expected peak down by a notch, not up. Le Bourget sits a few miles northeast of central Paris, and its readings can run cooler than the dense, concrete-baked arrondissements where the human cost actually lands.

That last point matters editorially. A reader in the 19th arrondissement may have felt 38 in the shade last week. The contract does not care. It resolves off one airport thermometer, on one day, and the mechanics of how prediction markets decide a winner are unforgiving about that kind of specificity.

Why a temperature ladder is a useful object

Weather contracts are a strange corner of the prediction market world. They are not really about uncertainty over whether a thing will happen, the way a Senate race or a film opening might be. They are about resolving a continuous variable, the actual temperature, into a set of discrete yes/no buckets. The result is a probability distribution you can read off the screen, and once you understand how prediction market odds work, it is genuinely informative. The market is telling you, in real time, where the meteorological consensus sits.

Compare that with a poll. A poll asks people what they think will happen. A prediction market asks people to put money on it, and if a forecaster on the other side of the world has a better model, that model gets expressed in the price. The literature on why prediction markets tend to be accurate leans heavily on weather and sports for exactly this reason: clean resolution, plentiful data, and a strong informational backbone.

The catch, of course, is volume. The 24-hour turnover on this contract is modest, the kind of figure that suggests a few engaged traders rather than a wisdom-of-crowds tidal wave. Thin markets are noisier markets, and a single sharper forecaster can push a band several points without breaking a sweat. That is worth holding in mind before treating any one rung as gospel.

The climate context the contract cannot price

The contract resolves on 30 June. It does not resolve on the heatwave that follows, on the hospital admissions in Seine-Saint-Denis, on the question of whether French cooling infrastructure is keeping up. The Guardian piece is, fundamentally, a political and emotional document: a pregnant woman in the poorest département describing what it is like to feel abandoned by the state during a record event. The Polymarket ladder is a meteorological one.

Both are real. Both matter. The ladder will give you a precise read on what traders expect a specific thermometer to do on a specific day. It will not give you a read on whether 28°C at Le Bourget translates to something far more dangerous in a top-floor flat in Saint-Denis without shutters. The market measures what it measures, and pretending otherwise is the fastest way to misuse a prediction market.

For what it is worth, the band concentration tells you something climate-relevant on its own. A market in which the modal outcome for 30 June is sitting in the upper twenties, with single-digit probability on anything below 26°C, is a market that has internalised the new normal. Forty years ago that distribution would have looked different. iPredicta tracks these climate-adjacent contracts across Polymarket and the regulated venues because the slow drift in where these distributions sit is, in its own undramatic way, one of the more honest pieces of climate data the markets produce.

Frequently asked questions

How does the Paris temperature contract resolve?

It resolves off the highest temperature recorded on 30 June 2026 at the Paris-Le Bourget Airport weather station, as published by Wunderground. Each one-degree Celsius band is a separate yes/no contract, and only the band containing the day's peak pays out. The location matters: Le Bourget sits a few miles northeast of central Paris and can run a touch cooler than the dense urban core.

Does the market reflect the wider heatwave or just one day?

Just the one day, and just the one thermometer. The contract is silent on the multi-day heatwave the Guardian reported on, on the human cost in places like Seine-Saint-Denis, and on whatever extreme heat is forecast for next week. It is a narrow meteorological instrument, useful for what it measures and easy to misread if you stretch it beyond that.