Kalshi Weather Markets — Overnight Lows: How Last Night Settled (July 27)
Welcome to the noon check-in on our Kalshi weather markets log, the running experiment where we sell long-shot temperature outcomes to people who overpay for them and publish every fill, every pass, and every loss. One disclosure before anything else: Stokastic holds these positions. We are the seller on every contract described here, and we profit if these contracts expire worthless. Read everything below with that in mind.
Monday's honest headline: there was nothing to grade last night, and we placed nothing at the noon window. On a book where a winning trade pays between four cents and $1.18, that is not a dead edition. It is the method showing itself, because in this strategy the pass is a position too. The most instructive line on our whole ledger is still the one loss, and we will get to it below.
In Summary
- Last Night Settled Clean Because It Settled Empty. We carried no low-temp positions into the overnight, so the low book added zero settlements. Every one of our 21 filled positions to date has been a high-temp trade; Sunday's final tally on that side was 6-0 for +$4.71.
- The Noon Window Placed Nothing New. The feed was healthy and the board was live; nothing changed hands in this slot. We have 24 resting orders on file across the book.
- Tuesday's Low Board Has 14 Tails We Want, all asked at 5 or 6 cents, all inside the 1-6 cent band this strategy actually trades, with queues of 2 to 7 orders ahead of us.
- We Passed On Four Tickers for one reason: queue depth. Being 350th or 506th in line at the same price is not a trade, it is adverse selection with extra steps.
- The Live Record Is 20-1 On 21 Filled Positions And Still Down $2.30, and the calibration read says the edge is not yet established at n=24. Both facts stay on the page.
What The Noon Window Produced: Nothing, On Purpose
We work resting limit orders only, post-only, buying NO. Taking the offer loses here; the entire theoretical edge is in being the maker, which also keeps us on the right side of the fee schedule (our breakdown of Kalshi's fees covers why the maker side of that formula matters). So a session where nothing fills is not a malfunction. It means nobody crossed the spread to pay our price, and paying a worse price ourselves is the one thing the strategy forbids.
That patience only makes sense if you understand what one mistake costs, which is the next section, and it is the section this series will repeat every single day.
One Claim Costs About 36 Winning Premiums
This book is short volatility. At our standard size, a winning position pays somewhere between $0.04 and $1.18, and a losing one costs about $18. Run that arithmetic once: a single claim wipes out roughly 36 winning premiums. A red day that erases a green month is not this strategy breaking. It is the shape of this strategy working as designed.
Anyone tempted to copy these trades should sit with that paragraph before anything else on the page. A 20-1 record here does not mean what a 20-1 record means anywhere else, and a reader who mirrors the fills without pricing the tail is the outcome we care most about avoiding.
A Worked Example: One 6-Cent Tail
Take the top row of Tuesday's board, Miami's 73° to 74° bracket, and walk the whole trade through. A buyer is offering 6¢ for YES, which means the market prices about a 6% chance Miami's Tuesday morning low lands exactly in that two-degree window. We post a resting NO order at 94¢ and wait; at our standard size that puts $18.80 of collateral behind the position. If nobody meets our price, nothing happens, which is exactly what this noon window produced across the book. If the order fills and Miami's low settles anywhere outside 73-74°, the contracts expire worthless and we keep the premium; at this standard size, winning positions on our log have paid between $0.04 and $1.18. If the low lands inside the band, we lose the collateral: roughly an $18 hit, the 36-premiums event. Same arithmetic on every row; only the city and the band change.
Tuesday's Low Board: 14 Tails We Want, Four We Passed
Now the look-ahead, and a date note first, because low-temp contracts are easy to misread: tonight's overnight low prints Tuesday morning, so it settles as Tuesday, July 28's ticker. Everything below targets Tuesday's lows, with roughly 35 to 38 hours to close. The running-minimum guard, the observation check that tells us when a band is already in trouble, has nothing to say yet, because Tuesday's observation window has not opened and no Tuesday readings exist. The temperatures you can see on the boards right now belong to Monday's tickers, today's live context only: those running minimums range from 59°F in Boston and Seattle to an absurd 94°F in Phoenix, and none of them count toward the Tuesday contracts below.
The full candidate list, straight from the board:
| City | Band | YES Ask | Our NO Limit | Queue Ahead | Hrs to Close | Collateral | Static Edge |
|---|---|---|---|---|---|---|---|
| Miami | 73° to 74° | 6¢ | 94¢ | 2 | 35.5 | $18.80 | 3.5¢ |
| Austin | 69° to 70° | 6¢ | 94¢ | 2 | 36.5 | $18.80 | 3.5¢ |
| Dallas | 74° or below | 6¢ | 94¢ | 2 | 36.5 | $18.80 | 3.5¢ |
| Seattle | 55° to 56° | 6¢ | 94¢ | 2 | 38.5 | $18.80 | 3.5¢ |
| New Orleans | 75° to 76° | 6¢ | 94¢ | 2 | 36.5 | $18.80 | 3.5¢ |
| Minneapolis | 67° to 68° | 6¢ | 94¢ | 3 | 36.5 | $18.80 | 3.5¢ |
| Chicago | 72° or above | 6¢ | 94¢ | 4 | 36.5 | $18.80 | 3.5¢ |
| New York | 73° or above | 6¢ | 94¢ | 6 | 35.5 | $18.80 | 3.5¢ |
| Houston | 72° to 73° | 6¢ | 94¢ | 6 | 36.5 | $18.80 | 3.5¢ |
| Philadelphia | 65° to 66° | 6¢ | 94¢ | 7 | 35.5 | $18.80 | 3.5¢ |
| Denver | 72° or above | 5¢ | 95¢ | 2 | 37.5 | $19.00 | 2.51¢ |
| Boston | 60° to 61° | 5¢ | 95¢ | 2 | 35.5 | $19.00 | 2.51¢ |
| Atlanta | 69° to 70° | 5¢ | 95¢ | 2 | 35.5 | $19.00 | 2.51¢ |
| Oklahoma City | 70° or below | 5¢ | 95¢ | 2 | 36.5 | $19.00 | 2.51¢ |
Every ask on this list sits inside the 1-6 cent band we actually trade, and sizing is the same on all of them: equal minimum size, every position. The risk that can actually wipe this book is one air mass taking out many cities at once, and the answer to correlation is breadth across different climates, never more size on a favorite. Queue position matters as much as price on thin boards like these; our look at Kalshi liquidity across markets covers why.
The row worth staring at is Dallas, "74° or below." Most bands on this board can die early: once a city's running minimum drops through a bracket, the market is effectively decided hours before settlement. A floor band cannot. The running low only moves down, so "74° or below" stays winnable for the buyer until the final observation of the window, and the seller never gets a moment where the trade is mathematically safe. Dallas's "74° or below" and Oklahoma City's "70° or below" are the two tickets on this list flagged that way, which is exactly why they had better be priced like the long shots they are.
And the passes, which no one else publishes and which explain the method better than any fill:
| Ticker | Why We Passed |
|---|---|
| Chicago — KXLOWTCHI-26JUL28-T64 | Queue of 350 ahead at the price |
| Philadelphia — KXLOWTPHIL-26JUL28-T65 | Queue of 162 ahead |
| Boston — KXLOWTBOS-26JUL28-T56 | Queue of 506 ahead |
| Boston — KXLOWTBOS-26JUL28-B58.5 | Queue of 161 ahead |
Think about what a queue of 506 means when you are selling a 5-cent tail. You only get filled if the market chews through 505 orders at your price first, and the market mostly does that when the weather is moving against the sellers. The fill you finally get is the fill you did not want. This is the same reason our whole approach is maker-only, and it is the same logic as refusing a bad number at a sportsbook: the price you accept is the whole game. That discipline is not unique to weather, and it is exactly the fair-vs-offered comparison the top bets screen runs on every sportsbook market all day.
The Record: 20-1, And Still Down $2.30
The live record through Sunday: 20-1 on 21 filled positions, $2.30 realized loss. Three resting orders never filled, and per the counting rule this log lives by, wins and losses count filled positions only; never-filled orders are reported separately and never as wins. Nothing is awaiting settlement.
| Date | W | L | P&L | Unfilled |
|---|---|---|---|---|
| July 24 | 7 | 1 | -$10.19 | 1 |
| July 25 | 7 | 0 | +$3.18 | 1 |
| July 26 | 6 | 0 | +$4.71 | 1 |
Now the promised loss, because it is the best teaching trade on the ledger. On July 24 we sold NO at 91 cents on Washington, D.C.'s 81-82°F high-temp bracket. That is a 9-cent tail, outside the band this strategy validates (2-6 cents on the high-temp book), and D.C. landed in the bracket. Cost: $18.23, on a day when our seven winners combined could not cover it. That is the 36-premium arithmetic from earlier, live and in person: one claim turned a 7-1 day into a $10.19 loser and still has the whole book underwater three days later. The lesson was not "weather is random." The lesson was that the band exists for a reason, and the one time we stepped outside it, the market collected.
Is This Actually Working? Honestly: Unproven
The calibration read, quoted straight because it was written to be quoted: n=24 settled tails. Implied 3.7% vs realized 4.2%, but the 95% interval on the realized rate is 0.7-20.2% and it spans the implied rate, so this is NOT significant in either direction. It takes roughly 1,100 settled tails to resolve an edge this small. Judge the strategy on this number, never on a night's P&L.
Inside the 1-6 cent band we actually trade, the sample is 22 tails and zero claims against an implied 3.3%. Twenty-two is not evidence; it is a start.
What made us run the test at all is the backtest, and the backtest is not the record; the two never blend. Fill-adjusted over past seasons, the pooled low-temp book showed a 33% fill rate and a +1.26 cent edge at 99% confidence, with spring the strongest read (24% fill, +1.72 cents, 99%) and winter the weakest (51% fill, +0.81 cents, 86%). The measured seasonal law says the fill-adjusted edge roughly halves from November through March on both books, so a weak winter would be the expected shape, not a broken strategy. And one range is off-limits entirely: the 6-20 cent low-temp tails only backtested well in spring, and the pooled number hides a bad winter, so we do not trade it, period. If the mechanics of these markets are new to you, start with how prediction markets work and our wider look at Kalshi strategies — this series is one deliberately narrow application of those ideas.
The Same Discipline, Cheaper Lessons
Strip the weather out and this whole series is three habits: know the fair price, refuse to pay worse than it, and count honestly afterward. Those habits transfer. On the sportsbook side, OddsShopper's odds screen shows you the market's fair number next to every book's offered odds, which is the identical comparison we run on these 5-cent tails, just with far more liquid markets on the other side. The maker-vs-taker lens transfers too: the Liquidity Tool reads where sharp money is sitting on exchanges and prediction markets, the same order-book thinking behind every queue pass above. If this log's method appeals to you more than its instrument does, those screens are where it pays fastest, and my current betting card lives on my expert page on Tails.
Take 20% off OddsShopper Pro with code KALSHIWX20 and put the fair-price habit to work on the bets you were already making, including the ones the price says to skip.
Kalshi Weather Markets FAQ
What are Kalshi weather markets? They are event contracts on a CFTC-regulated exchange that pay $1 if a city's official high or low temperature lands in a stated band and $0 if it does not, which is why the Miami worked example above nets pennies when we win and about $18 when we lose. Prices between 0 and 100 cents read directly as implied probabilities: that 6¢ YES ask is the market saying 6%. If the mechanics are new, our guide to how to bet on Kalshi covers order types, settlement, and fees.
How is this different from sports betting? There is no bookmaker taking the other side; you trade against other participants on an order book, which is why maker-vs-taker position matters so much here. The full comparison lives in prediction markets vs. sports betting.
Why sell the tails instead of buying them? Our working hypothesis, supported so far only by backtest data, is that retail buyers overpay for long-shot temperature outcomes the way they overpay for lottery-ticket parlays. Selling at 94-95¢ collects that premium, and holding to settlement avoids paying the spread twice. Whether that premium is real is exactly what the calibration section above is measuring.
Is the strategy profitable? Not yet, and we will not dress that up: the live log is 20-1 on filled positions and down $2.30, and at n=24 the sample cannot establish an edge in either direction. This page exists to find out in public.
The Fine Print
Stokastic holds these positions. We are the seller on every contract described on this page, and we profit if these contracts expire worthless. Kalshi weather contracts are CFTC-regulated event derivatives traded on a designated contract market, not sportsbook wagers, and they can lose their full value. They are available to adults 18 and older where Kalshi operates (legality varies by state and instrument), and the risk of loss is real; on the side we trade, each individual loss is large. This is an open research log of a strategy we have not yet proven. Nothing here is trading advice.
Tonight
The 6:30 edition will show whether any of these 14 orders earn a fill before Tuesday's observation window opens, and the running-minimum guard finally gets a vote once real temperatures start printing. If nothing fills again, you will read that too. On a book where one claim costs 36 premiums, the days when nothing happens are not the days the strategy is failing. They are the days it is holding its price.
