NFL WEEK 1

Model vs. Kalshi on all 16 openers

Week 1 markets live now

FOMC SEPT 16

Kalshi, Polymarket and futures disagree on the Fed

Decision Wed 2pm ET

NFL EDGES

Live game edges — model vs. Kalshi, ranked

Gridiron Edge · 5-min refresh

COMBO EDGE

Correlation-aware Kalshi combos — free

SILVER EDGE

Options vs. Kalshi on weekly silver brackets

Home/Open Data

Open Data Catalog

12 datasets, free to use with attribution. Market-implied probabilities across macro, politics and sports contracts; a composite read on the US economy; a Monte Carlo tournament simulation; and the settled, graded record of how our own engines have done. Every one is available as JSON and as CSV, with no key and no signup.

Licence and attribution

Everything below is published under Creative Commons Attribution 4.0. Commercial use is fine, redistribution is fine, building a product on it is fine. Name the source and link back.

Suggested credit line: Data: PredictionMarketsPicks (predictionmarketspicks.com)

Datasets

Market-implied probability that the NBER declares a US recession, read from live Kalshi binary contract prices. One row per listed contract, each carrying the implied probability, 24-hour volume and the snapshot time it was read.

Measures
  • Implied probability of an NBER-declared US recession
  • 24-hour contract volume
  • Snapshot timestamp (UTC)
Dataset pageJSONCSVCoverage: 2025/2027

Market-implied probability distribution for US CPI year-over-year prints, derived from Kalshi threshold contracts. Reading the price across every listed threshold gives a full implied CDF, and from it an implied median.

Measures
  • Implied probability that CPI year-over-year exceeds a listed threshold
  • 24-hour contract volume
  • Snapshot timestamp (UTC)
Dataset pageJSONCSVCoverage: 2025/2027

Market-implied probability distribution of the S&P 500 year-end closing level, derived from Kalshi bracket contracts. Exactly one bracket resolves, so the prices across brackets read directly as a probability distribution.

Measures
  • Implied probability that the S&P 500 closes within a listed bracket at year end
  • 24-hour contract volume
  • Snapshot timestamp (UTC)
Dataset pageJSONCSVCoverage: 2026

US Government Shutdown Probability

Updates: Every 5 minutes while a deadline contract is listed

Market-implied probability of a US federal government shutdown at each funding deadline Kalshi lists a contract for. Active in the weeks around a deadline; between deadlines the series is intentionally sparse rather than back-filled.

Measures
  • Implied probability of a US federal government shutdown at a listed funding deadline
  • 24-hour contract volume
  • Snapshot timestamp (UTC)
Dataset pageJSONCSVCoverage: 2025/2027

A single 0–100 score for US macroeconomic health, recomputed daily from eleven public FRED series across six categories — growth, labor, inflation, rates, liquidity and sentiment — together with the regime band the score falls in and each category sub-score.

Measures
  • Composite macro health score (0–100)
  • Regime classification (Recession Watch, Risk Off, Neutral, Risk On, Expansion)
  • Per-category sub-score for growth, labor, inflation, rates, liquidity and sentiment
Dataset pageJSONCSVCoverage: 2025/2027

A daily series of implied probabilities for each outcome of a named FOMC meeting on three venues: Kalshi, the Polymarket international order book, and an inversion of CME fed funds futures. Only same-basis rows are ever subtracted — Kalshi asks a cumulative question about the target range following a meeting while the other two price the move at that meeting, so the questions coincide for the next meeting and diverge by pure basis after it.

Measures
  • Daily implied probability of each FOMC rate outcome, per venue
  • Venue label
  • Question basis (per-meeting or cumulative)
  • Same-basis cross-venue spread in percentage points
Dataset pageJSONCSVCoverage: 2026/2027

The settled, graded record for every PMP predictive tool published with public visibility. Each signal is written down before resolution and graded against the actual settlement, so the losing engines appear alongside the winning ones. Scored on the all-time window.

Measures
  • Brier score
  • Hit rate with a Wilson 95% confidence interval
  • Return on investment at par
  • Maximum drawdown
  • Count of decided signals
Dataset pageJSONCSVCoverage: 2025/2027

Live implied probability for the 2026 US midterm races and chamber-control questions we track, read from contract prices. Every row names the venue it came from; any Polymarket row is the international order book, the only Polymarket venue we ingest.

Measures
  • Implied probability of the named election outcome
  • Source venue
  • 24-hour contract volume
  • Snapshot timestamp (UTC)
Dataset pageJSONCSVCoverage: 2025/2027

Live implied probability for 2028 US presidential nomination and general-election outcomes, read from contract prices across both platforms we cover. Every row names the venue it came from; Polymarket rows are the international order book, the only Polymarket venue we ingest.

Measures
  • Implied probability of the named candidate or party outcome
  • Source venue
  • 24-hour contract volume
  • Snapshot timestamp (UTC)
Dataset pageJSONCSVCoverage: 2025/2029

Champion probability for all 48 teams from The 7 Oracles 10,000-iteration Monte Carlo tournament simulation, with the probability of reaching the quarter-final, semi-final and final, plus each team’s group. Model output, not market prices — the source field reads "simulation" on every row.

Measures
  • Modelled probability of winning the tournament
  • Modelled probability of reaching the quarter-final, semi-final and final
  • Group assignment and model rank
Dataset pageJSONCSVCoverage: 2026

Cross-Venue Sports Contract Board

Updates: Every 15 minutes

Contract-level implied probabilities for the sports markets our cross-venue scanner tracks, each row labelled with the venue it was read from and carrying 24-hour volume. This is the board the Kalshi-versus-Polymarket comparison is computed from, published as rows rather than as a derived gap.

Measures
  • Implied probability of the named outcome on the venue given in the source field
  • Source venue
  • 24-hour contract volume
  • Snapshot timestamp (UTC)
Dataset pageJSONCSVCoverage: 2026/2027

Premier League Fixture Contract Board

Updates: Every 15 minutes

Implied probabilities for the Premier League fixtures listed as event contracts, one row per fixture with the venue it was read from and 24-hour volume. Fixture-level, not aggregated to a season table.

Measures
  • Implied probability of the named fixture outcome on the venue given in the source field
  • Source venue
  • 24-hour contract volume
  • Snapshot timestamp (UTC)
Dataset pageJSONCSVCoverage: 2026/2027

Using the API

Append ?format=csv to any endpoint for the CSV form. Responses are CORS-open and cached for about a minute, and the free tier is rate limited per IP rather than by key. Read the snapshot timestamp on each row to tell a reader how fresh a number is.

Agent-facing documentation lives at /llms.txt, /llms-full.txt and /skill.md. Agents that prefer to call the analysis directly can use our MCP server.

Frequently Asked Questions

What licence is this data published under?

Creative Commons Attribution 4.0 International (CC BY 4.0). You can use it commercially, redistribute it, and build on it, including in a product you charge for. The one condition is attribution: name PredictionMarketsPicks and link back to the page the data came from. That is the whole deal — free data, credit where it came from.

How do I attribute it?

A line such as "Data: PredictionMarketsPicks (predictionmarketspicks.com)" with a working link to the dataset page listed here. If you are publishing a chart or a table, put it in the caption. The free tier of the API also returns an X-Attribution header carrying the same string, so an automated pipeline can pick it up without hard-coding anything.

Do I need an API key?

No. Every endpoint listed here is open, needs no key and no signup, and allows cross-origin requests. The free tier is rate limited per IP, which is generous enough for a dashboard or a research script. Paid keys exist for high-volume use and raise the ceiling; they do not unlock different data.

What does an implied probability actually mean here?

On these exchanges a contract settles at $1 if the event happens and $0 if it does not, so its price in cents is the market’s own probability estimate — a 25c contract implies 25%. No odds conversion and no vig to strip out. Where a dataset carries a model number instead of a market price, the description says so explicitly.

How fresh is each dataset?

It varies by dataset and the cadence is listed against each entry below, from every 5 minutes for the live market trackers to nightly for the simulation and daily for the composite scores. Every row carries its own snapshot timestamp, so never infer freshness from when you fetched it — read the timestamp on the row.

Are these datasets ever revised after publication?

Market snapshots are not — a row records what a contract was priced at when we read it, and that does not change afterwards. The graded scorecards do move, because a signal that is still open gets a settled result later; that is a new grade appearing, not an old one being edited.

What is deliberately not published here?

Model internals and any licensed upstream data. What we publish are derived outputs: implied probabilities, composite scores, edges, tiers and graded results. Datasets that have no public landing page, or whose endpoint currently returns no rows, are left out of this catalog rather than listed as broken links.