Learn · Weather

How to trade weather on prediction markets

Weather is one of the most actively traded categories on Kalshi, and the contracts are usually about a single number: "Will the high in New York be above 75°F today?" Unlike most categories, weather has a genuinely clean structural edge — because the outcome is settled by an official measurement that is forecast well and that locks in as the day goes on. This is a plain-English guide to what those bracket prices mean, where the edge actually comes from, and how the free WX Lookup tool on this site shows the live observations, the forecast, and the nearest bracket behind every market. Everything here is paper trading. No real money, ever.

1. A bracket price is an implied probability

A weather market isn't "will it be nice out" — it's a yes/no question about a specific official number with a deadline of the end of the day: "High temperature in NYC above 75°F?" The YES contract trades between 0¢ and 100¢. If it's at 40¢, the market is saying there's roughly a 40% chance the day's official high finishes above that line. Buy YES at 40¢ and you collect $1 if it does (a 60¢ profit) or lose your 40¢ if it doesn't.

Most cities have a ladder of brackets for the same day — above 75°, 74–75°, 73–74°, and so on — and the prices across the ladder add up to a probability distribution over where the high will land. The number that settles it isn't a vibe: it's the official high (and low) the National Weather Service records for that station — what we call the governing number.

2. The edge is the gap between your number and the bracket's

The edge framing is the same as anywhere: if the bracket implies a 40% chance and your own estimate is 55%, that 15-point gap is your edge — provided your number is genuinely better calibrated than the market's. Bet that consistently and you profit on average, even while losing plenty of individual days.

Why weather is different from crypto: the edge is structural and real.

Short-horizon crypto is close to a coin flip, so we don't ship a price predictor. Weather is the opposite case. Two structural facts give an honest edge: (1) the high is forecast well — professional numerical models predict tomorrow's temperature far better than chance — and (2) the day resolves in slow motion. As the afternoon unfolds, the running high is observed in real time, and once the day's peak has already happened and can't be beaten, the outcome is partly or fully decided before the market closes. A bracket that's already physically impossible — or already certain — but still trading away from 0¢/100¢ is the cleanest edge in the whole library.

3. A worked example: the WX Lookup board

The WX Lookup tool is the honest, real-data version of this for the 20 U.S. cities Kalshi writes daily temperature markets on — New York, Chicago, Phoenix, Miami, Denver, Los Angeles, and more. It's built on free, public National Weather Service data (live observations plus the official CF6 climate report each station files), and it gives you context, not a crystal ball:

The governing number, front and centre. Every panel shows the official high/low that will actually settle the market — not a generic app temperature — so you're judging the same number Kalshi judges.
Live observation vs. remaining-day forecast. The board shows the latest observed temperature alongside the forecast for the rest of the day, so you can see how much room is left for the high to move.
Locked highs and lows. When the day's high is already in and can't be beaten, the panel marks it locked — that's the moment a bracket's outcome is settled even if the market is still trading on it.
The nearest live bracket, with distance. For each city the board pulls the closest real Kalshi temperature market and shows the distance from the current/forecast number to the threshold next to its YES price — the raw material for spotting a gap.
Source and provenance on everything. Each number is labelled with where it came from and how precise it is, with a staleness badge — no fabricated readings, ever.

The lesson mirrors the rest of the site: surface the real, settling number, be honest about how much of the day is still uncertain, and only bet a gap when you have a calibrated reason it's real.

4. Turn a view into a bot

You don't have to watch the tape all day. On this site you can wire a view into a paper-trading bot that runs 24/7: the weather anchor lets a strategy fire only when a city's forecast or locked high sits in the zone you care about — opt-in, so it never touches your other bots. Start from the weather build page, pick your gates, and the bot backtests, then trades live on the public leaderboard in paper money so you can watch the idea prove out (or not) in the open.

5. Honest caveats

The forecast is good, not perfect. Most of the time the market already reflects the best available forecast — assume the gap you "see" is priced until you can explain why it isn't (a model disagreement, a stale quote, a bracket the crowd is slow to re-price as the day locks in).
The governing number is specific. Markets settle on the official station high/low, which can differ from the temperature on your phone. Trade the number that settles the contract, not the one out your window.
Observations are public reference data. NWS feeds are free and authoritative for settlement, but they update on their own cadence — check the staleness badge before acting on a "locked" read.
This is a sandbox. Everything here is simulated paper trading — a place to test ideas, not financial advice and not real-money order placement.

Open WX Lookup → Build a weather bot See the signal library

New to prediction markets? Start with how it works for the mechanics of brackets, pricing, and settlement. Trading a different category? See the crypto and sports primers.