Weather prediction markets let traders buy and sell contracts based on future weather outcomes. Instead of betting on sports or elections, you're trading on whether tomorrow's high temperature in Chicago will exceed 45°F, or whether NYC will get more than 0.5 inches of rain this week.
How Weather Contracts Work
A weather prediction contract is a binary event market. Each contract resolves to either YES (the event happened) or NO (it didn't). For example:
- "Will NYC's high temperature exceed 50°F on March 25?"
- If the answer is YES, the contract pays out $1.00
- If NO, it pays $0.00
- The current trading price (say $0.72) represents the market's implied probability (72%)
Where to Trade Weather Markets
### Kalshi Kalshi is a CFTC-regulated exchange based in New York. They offer daily weather contracts for major US cities covering temperature highs/lows, precipitation, and snowfall. Contracts are denominated in USD and settle based on official NWS (National Weather Service) observed data.
### Polymarket Polymarket is a decentralized prediction market that occasionally lists weather-related contracts. It uses a CLOB (Central Limit Order Book) system and settles in USDC. Weather markets on Polymarket tend to be less frequent but can cover broader events like seasonal records.
Why Weather Markets Are Interesting
Unlike political or sports markets, weather contracts have a unique advantage: numerical weather prediction models provide independent probability estimates. The GFS (Global Forecast System) and ECMWF (European model) generate probabilistic forecasts that can be directly compared against market prices.
This means you can systematically identify contracts where the market disagrees with the models — what traders call "edge."
What Is Edge in Weather Trading?
Edge is the difference between what forecast models predict and what the market prices in. If GFS and ECMWF both suggest a 90% chance of NYC hitting 50°F, but the Kalshi contract is trading at $0.72 (72%), that's an 18-point edge.
Celsi calculates edge scores by: 1. Fetching current contract prices from Kalshi and Polymarket 2. Querying GFS and ECMWF ensemble forecasts 3. Calculating model consensus probability 4. Scoring the divergence between market and models 5. Adjusting for historical model bias in each city
Getting Started
The simplest approach is to focus on contracts where model consensus strongly disagrees with market pricing — typically edge scores above 60 on Celsi's scanner. Start with paper trading to build intuition before risking real capital.