A practical guide to this Kalshi category — what the markets look like, how settlement works, where the edge actually is, and the resources for building a strategy. Live current markets and FAQ below.
Kalshi lists weather prediction markets for major US cities covering temperature (high, low), precipitation (rain, snow), wind, and hurricane outcomes. Each contract resolves to $1 if the event happens and $0 if not — so the price between 1¢ and 99¢ represents the market's implied probability of the outcome.
The reason weather is one of the more reliable categories for systematic trading: public forecasts (NWS, Open-Meteo) are the same data the market uses to price contracts. When you can pull the same forecast as everyone else, the edge comes from disciplined risk management, not from secret information. A 2°F forecast cushion over the temperature threshold is typically enough to beat Kalshi's 1.75¢ per-contract fee.
Every weather market settles on a specific NWS station of record — not the city's general forecast, not your airport, not your local thermometer. The Boston temperature market, for example, settles on the Logan Airport KBOS station reading at the close time printed on the market. Misreading the station of record is the #1 source of "I lost a trade I thought I'd win" stories — always check the settlement language before sizing.
Daily high contracts settle on the highest temperature observed at the station between 00:01 and 23:59 local time. Daily low contracts settle on the lowest. Rain markets settle on cumulative measurable precipitation at the station for the calendar day.
Across weather markets, the 88-95¢ NO band has historically returned 1.5-2.4% per trade after fees when the forecast cushion exceeds 2°F. The strategy is to identify markets where the forecast clearly favors one side, restrict to the favorite-band price range, and accept the asymmetric payoff (risk 88¢ to make 12¢, but win ~91% of the time).
The risk: one loss costs you roughly 7 wins to recover. Sizing must reflect that — half-Kelly or smaller on every trade.
| Market | NO ask | Spread | 24h vol |
|---|---|---|---|
| Will it rain in New York City on Sep 13, 2026? | 1¢ | 1¢ | $356 |
| Will the maximum temperature be >83° on Sep 13, 2026? | 97¢ | 2¢ | $295 |
| Will it rain in Trenton on Sep 13, 2026? | 1¢ | 1¢ | $207 |
| Will it rain in Oklahoma City on Sep 13, 2026? | 1¢ | 1¢ | $179 |
| Will the maximum temperature be 82-83° on Sep 13, 2026? | 94¢ | 2¢ | $161 |
| Will it rain in Houston on Sep 13, 2026? | 50¢ | 1¢ | $153 |
| Will the maximum temperature be 80-81° on Sep 13, 2026? | 69¢ | 2¢ | $145 |
| Will the maximum temperature be 78-79° on Sep 13, 2026? | 45¢ | 2¢ | $141 |
| Will it rain in Washington DC on Sep 13, 2026? | 1¢ | 1¢ | $123 |
| Will the maximum temperature be 72-73° on Sep 13, 2026? | 100¢ | 99¢ | $107 |
| Will it rain in Newark on Sep 13, 2026? | 1¢ | 1¢ | $100 |
| Will the maximum temperature be <78° on Sep 13, 2026? | 64¢ | 1¢ | $94 |
Each market resolves to $1 if the event (e.g., 'high temperature above 75°F in Boston on May 28') happens, and $0 if not. Prices between 1¢ and 99¢ represent the implied probability. They settle on a specific NWS weather station of record — check the market language carefully, since the station may differ from what you'd expect.
Empirically: 88-95¢ NO contracts on markets where the public NWS or Open-Meteo forecast has at least a 2°F cushion over the threshold. Historically returns 1.5-2.4% per trade after fees. The win rate is high (~88-92%) but losses are large relative to gains, so sizing must be conservative.
Read the market's settlement language — it always names the station explicitly (e.g., 'KBOS Logan Airport' for Boston). Don't assume it's the city center or the nearest airport to where you live.
High-population-center temperature markets (NYC, Chicago, LA) typically have $2,000-$20,000 in daily volume, which is enough for retail-size positions. Smaller cities and rain/snow markets often have <$500 in daily volume, which limits practical position size.
Analyze any of these markets in one click. The AI reads live prices, your positions, and external signals — and answers in plain English.
Try Prediction Pilot — 250k free credits