⛅ Weather Bot
Compares real forecast data to Polymarket weather contract prices. Signals when they diverge.
What it does
Polymarket runs thousands of weather contracts every day — will the high in Chicago be above 55F tomorrow, will it rain in London, etc. Our Weather Bot aggregates real meteorological forecast data and compares it to what Polymarket is currently pricing. When the forecast says one thing and the market says another, you get a signal.
Data sources
The model pulls from multiple forecast providers to build a consensus view:
- NOAA (GFS) — US government global forecast system, updated every 6 hours
- ECMWF — European model, widely considered the most accurate for medium-range forecasts
- NAM / HRRR — Higher resolution regional models for short-range US forecasts
- Ensemble averaging — Multiple model runs weighted by recent accuracy per region
How it calculates edge
The bot converts forecast data into a probability for each weather outcome, then compares that to the current Polymarket price. The difference is the "edge" — for example, if the model says 75% chance of the high exceeding 55F but Polymarket is pricing it at 41 cents ($0.41 = 41% implied probability), the model edge is +34%.
Signals are only fired when the edge exceeds a minimum threshold and the conviction score is high enough. This filters out marginal opportunities and markets with too little liquidity.
What you receive
Each alert includes the city, the specific weather market, the direction (YES/NO), the model's calculated edge percentage, a conviction score from 1-10, and the current Polymarket price.
Performance notes
Weather signals have shown the strongest alpha in the 10-30% edge bucket — markets where the forecast clearly disagrees with the odds but not so dramatically that the market has already corrected. The 24-48 hour forecast window is where the model performs best. Beyond that, forecast accuracy drops and so does our edge.