A machine learning model predicts the most likely 2-degree temperature bucket, scores it STRONG BET, LEAN, or SKIP, and backs it with live atmospheric signals and a reversal-risk panel — updated throughout the day.
Built around one question: which bucket should I bet, and how confident should I be?
A machine learning model trained on historical NWS, AccuWeather, and atmospheric data predicts today's and tomorrow's most likely 2-degree Kalshi bucket. Each prediction is labeled STRONG BET, LEAN, or SKIP based on model confidence.
Solar irradiance, cloud cover, 850mb and 925mb temperatures, and surface wind direction update in real time throughout the day — not just the overnight forecast. See what the atmosphere is actually doing right now, not what was predicted at 5 AM.
A dedicated panel below each prediction scores how likely the current bucket pick is to shift before settlement — and whether conditions support a blow-past of all model forecasts. Backed by historical data from similar setups and updated throughout the day.
Every prediction is scored against the official daily high. WIN/MISS history is stored in a database and shown transparently — so you can see exactly how the model has performed over time, not just marketing claims.
Top-1 and Top-2 most likely buckets shown with confidence percentages, so you see the full probability picture — not just one call.
Between 10 AM and 3 PM local time, the dashboard automatically increases its refresh rate and signals that it's watching closely. Observations update every 2 minutes during this window.
Current market leaders, prices, and settlement results pulled directly — so you can compare the model's call against what the market is actually pricing.
Predictions refresh throughout the day as observed temperatures and atmospheric conditions change — not just once in the morning. The model triggers a fresh run when conditions run warmer or colder than expected.
Full ML predictions, atmospheric signals, and Kalshi bet signals for both New York City and Los Angeles — two of the most actively traded daily high markets.
NWS and AccuWeather forecasts are logged every update, bias-corrected against historical accuracy, and weighted into the ensemble — building a forecast audit trail you can actually inspect.
A glimpse of what's in the dashboard right now.
Live conditions updated every 15 minutes — what's actually happening, not just the overnight forecast.
Get the ML prediction, live signals, and reversal risk — every day.
Only $5/month or $50/year.
A data guy who traded Kalshi's NYC daily high. I wanted a real edge: not just a weather app, but a purpose-built system that tracks every forecast issuance, trains a model on historical outcomes, incorporates live atmospheric conditions, and tells me whether to bet — and how confidently. Built for me; shared because it keeps getting better and others find it useful.