An algorithmic trading desk built on a weather model.
Two automated strategies on Kalshi's event markets. The flagship turns a skill-weighted, multi-model weather ensemble into market-implied probabilities and trades only the edges that survive a closing-line-value test. It's probabilistic forecasting, expected value, calibration, and risk — the same toolkit behind financial modeling and valuation.
● Walkthrough only — fabricated market data. P&L is redacted throughout; this shows the method, not the book.
A multi-model weather ensemble forecasts tomorrow's NYC high as a distribution, gets mapped onto Kalshi's temperature brackets, and where the model disagrees with the market the desk places an RSA-signed order — then alerts Telegram.
Two strategies running 24/7 as services — the weather edge and an expiration-premium collector — with closing-line-value tracking and Brier calibration before any edge is believed. P&L stays private; the method is the point.
