Weather Market NO-Side Overpricing Arbitrage

How It Works This strategy compares Polymarket's implied probability for temperature-threshold markets against a Gaussian probability model built from live METAR (aviation weather observation) data and Open-Meteo forecast data. When the market prices the YES side too high relative to the meteorological model's estimate, the bot takes the NO position.

Entry Conditions Entry is triggered when the market's implied YES probability exceeds the Gaussian model's computed probability by a defined threshold, indicating the NO side is underpriced. The model incorporates both current observations (METAR) and short-term forecasts (Open-Meteo) to construct a probability distribution around the temperature outcome.

Exit Conditions Positions are exited at market resolution. As a paper-trading bot, it does not specify dynamic exit rules, but the structure implies holding to expiry given the binary nature of temperature markets.

Verified Results No verified on-chain results are provided; the repo explicitly operates in paper-trading mode, limiting confidence in real-world PnL claims.

Implementation Notes & Risks The edge relies on the accuracy and timeliness of METAR and Open-Meteo data feeds. Key risks include model error (Gaussian assumptions may not capture tails well), data latency, and the possibility that other sophisticated weather traders have already closed mispricings. The paper-trading constraint means live slippage and liquidity costs are unaccounted for.