Quantitative edge for Kalshi weather traders. Eight shadow fair-value engines, Platt-calibrated, ticking every five minutes across 67 stations.
Most traders eyeball a forecast app and guess. Market makers eat the spread, take-off trades go unanswered, and strike probabilities drift from reality within minutes of the next model run. You need quant infrastructure, not another weather widget.
Consumer weather apps refresh once an hour. Kalshi strikes move every minute. By the time you see a change, the book has too.
NWS probabilities aren't book-calibrated. Raw GFS/HRRR output isn't either. Without a calibrated fair value, every trade is a coin flip with a spread on top.
You cross the spread because you can't rank edge. A 3¢ edge looks the same as a 0.5¢ edge in the UI. Result: negative expected value, one trade at a time.
Live shadow bot results. Not a backtest — actual orders on Kalshi with real P&L. Methodology in the docs.
An ensemble of shadow fair-value models, Platt-calibrated against realized outcomes, running in parallel across every weather series Kalshi lists.
GFS, HRRR, ECMWF, NAM, GEFS ensemble — blended with live METAR and PWS obs on a 5-minute tick.
Eight shadow FV models produce independent strike probabilities; Platt calibration maps them to bookable prices.
Edge-ranked trade tickets on every Kalshi weather contract. Taker-aware, Kelly-sized, ready to execute.
Trades temp / rain / wind contracts between jobs. Wants edge without a PhD.
Runs size. Wants raw fair values, historical calibration, and an API.
From DFS and sportsbooks. Weather is his next +EV venue.
14-day free trial on paid tiers. Cancel any time.
14 days free. No credit card for the Free tier. Cancel in two clicks.