Strike Ladders as Distribution Labels: CPI, Fed, Temperature — and Crypto Grids

Most prediction-market data guides talk about binary questions: did the team win, did the window close up. A strike ladder asks a numeric question — what will the CPI print be, where will the Fed set rates, how hot will it get in New York today — and answers it with a row of binary contracts, one per range. Priced together they are not a probability but a distribution, and a dataset built from them has a richer label than Yes or No. This guide explains what a ladder is on each venue, how to turn it into distribution labels and cross-strike features, what happens in the book on a release day, and why the hourly Bitcoin ladder on Kalshi is the same object with a faster clock.

Everything below runs on the CryptoStruct archive: every message Kalshi, Polymarket and 35+ crypto venues publish — every Level-2 snapshot and update at full depth, every trade, every contract, every day since we added the venue — captured co-located with nanosecond venue and receive timestamps and a gap-audited event-id chain, in one normalized schema. It is the same capture our own trading engine and enterprise feeds run on, sold as €1 day bundles you buy as a guest in the Data Shop with instant download — no subscription, no minimum, no sales call. Free full-day samples let you run every command in this guide before paying.

What a strike ladder is

A ladder is a family of contracts over one event that partitions the outcome: on Kalshi the CPI family lists contracts such as KXCPI-26OCT-T1.0 (inflation above a threshold) for each strike of a release, the Fed funds rate family a contract per rate level after a meeting, and a temperature family such as highest temperature in NYC a contract per degree band (KXHIGHNY-26OCT23-B63.5, below 63.5 °F). Polymarket lists the same structure per city and day in its daily high-temperature markets — 104 city series as of 2026-10 — with one market per band. Some ladders are cumulative (above or below a strike, as the T and B suffixes say), some are bucketed (between two values); both settle exactly one way per strike, and both can be converted into each other by differencing or summing along the ladder.

Every rung is its own contract file in the series-day bundle: its own order book, its own trades, its own settlement. A CPI release day on Kalshi is therefore not one market but a dozen or more, quoted side by side, and the series page's "Contract anatomy" block shows what share of the family's contracts carry a strike — on a ladder family it is close to all of them.

From one-hot rungs to distribution labels

Because exactly one rung settles in the money, the label of a ladder is one-hot over its rungs — a categorical label with an ordering. That ordering is what makes ladders more useful than a stack of unrelated binaries. From the one-hot label you can derive the realised value (the rung that settled), and from the ladder's prices at any timestamp you can derive the market's implied distribution: the probability mass per bucket, its mean (the market's expected print), its spread, its skew and how much mass sits in the tails. Models can then be trained on richer targets than a direction — the distance between the implied mean and the realised rung, the calibration of the implied distribution across releases, or the change of the distribution between two timestamps.

Release-day anatomy

Economic ladders have a schedule, and the schedule shapes the file. A CPI release is published at 08:30 Eastern on a known date; the Fed decision at 14:00 Eastern on a meeting day. In the days before, the ladder is quoted thinly and reprices on forecasts, prior releases and commentary; the implied distribution narrows as the date approaches. At the print, every rung reprices within seconds — the in-the-money rung runs towards 0.99, the others towards 0.01, and the order in which the market makers pull and re-place their quotes across the ladder is visible message by message. The series page's hour profile shows the UTC hour in which the family's contracts are listed to settle; the day file shows what happened in the minutes around it, nanosecond-stamped and recorded co-located with the venue.

A release-day study therefore has three windows: the pre-release distribution (features), the print (an exogenous event you join by timestamp from the statistical agency's release time), and the settlement (the label). The election-markets guide makes the same point about polls — align by the second against the book, never by calendar day against a close.

Temperature ladders: one per city, one per day

Daily temperature markets are the most regular ladders in the archive. Polymarket lists a high-temperature and a low-temperature ladder for each of 104 cities as of 2026-10, every day, with one market per degree band; Kalshi lists comparable families for major US cities. Each day is a fresh ladder that settles against the official weather station reading, so a single city family yields one distribution example per day — hundreds per year — and the weather forecast is an exogenous feature with a clean timestamp. The series pages show how many contracts of such a family trade per day and, where the sample allows, how often the band the opening price favoured settled in the money; read that figure as a property of the day's most traded bands, labelled implied from the last recorded print.

Temperature ladders also expose the two market conventions side by side: Polymarket markets are mostly bucketed ("between 61 and 62 °F"), Kalshi contracts mostly cumulative ("below 63.5 °F"). A parser that converts both into one cumulative distribution per timestamp makes the two venues comparable — the temperature topic hub lists the families on each.

Cross-strike features

A ladder constrains itself. In a cumulative ladder the price of "above 1.0" must be at least the price of "above 1.1"; in a bucketed ladder the prices must sum to about one. Violations of those constraints in the book are either stale quotes or opportunities, and the depth behind each rung tells you which — a feature no single contract can provide. Further cross-strike features follow: the spread of each rung against the spread of its neighbours (market makers quote the body tighter than the tails), the depth profile across the ladder, how far from the implied mean the turnover concentrates, and how the whole curve shifts rather than how one price moves. The order-book dataset guide builds the per-contract grid these features are computed from; a ladder is that grid stacked across strikes.

The same shape with a faster clock: crypto ladders

The Kalshi Bitcoin above/below family is a strike ladder that settles every hour: a row of contracts across price levels, one in the money per hour, roughly 200 strikes listed at a time. It is the same object as the CPI ladder with a clock 700 times faster, and it is the bridge between this guide and the up/down training-data guide: the crypto grids give you label counts in the hundreds of thousands per year, the economic ladders give you a handful of releases, and the hourly ladder sits in between with 24 distributions a day. Features developed on the fast ladder — curve shifts, tail mass, body-versus-tail spreads — transfer to the slow ones; the slow ones add the release-day structure the fast ladder never has.

Splits: by release

Split by release or event day, never at random. Every rung of one ladder settles on the same print, and consecutive releases of one series share the macro regime; a random split over rungs leaks the print, a random split over days leaks the regime. Walk forward over releases — train on the ladders that have settled, validate on the next release, roll — and for daily temperature ladders treat seasons as regimes and cities as separate but related series. Report calibration of the implied distribution per release block, not pooled.

Data plan

1. Pick a ladder family by its clock: monthly releases for release-day studies, daily temperature ladders for many examples with an exogenous forecast, the hourly Bitcoin ladder for volume.
2. Download the free sample contract of the series page and the full-day samples; write the ladder parser first — strike from the identifier, cumulative versus bucketed, one distribution per timestamp.
3. Buy release days as ranges with /shop/bundles?bundle=<series-key>&d=<from>..<to> so the pre-release quoting is in the set, not only the print.
4. Derive the one-hot label from the settlement and the implied distribution from the book; audit the ladder constraints as a parser check.
5. Walk forward over releases, count ladders as examples, and report calibration per block.

Coverage

The Economics and Climate category hubs list every ladder family; the CPI, Fed and temperature topic hubs group them. Everything is sold the same way: €1 per series-day, no subscription, no minimum, no account.

Limitations

The statistics on series pages cover the ten most traded contracts per day — the bundle holds every rung, the statistics do not, and on a ladder the top-10 leans towards the rungs near the implied mean. Registry counts quoted here are as of 2026-10. Polymarket files between 2025-09-16 and 2026-01-08 are affected by an adapter fault that left about half of them without order-book deltas (trades are complete); those days are sold as-is and marked in the shop. Kalshi trades between 2026-02-26 and 2026-03-10 were recorded twice by the first recorder — de-duplicate on the trade id. All venues have a late start on 2026-06-17. Outcomes in our statistics are implied from the last recorded print, never the venue's official settlement, which you should join as the authoritative label; the realised print itself (the CPI figure, the temperature reading) is not part of the market data. Venue APIs expose some historical market and price data, but not the tick-by-tick book path — check the current docs. Nothing here is betting or investment advice.