# Dataset manifest for MKT-013 — what will be the largest company in the world
# by December 2026.
# Schema: schemas/data-manifest.schema.json (EPIC §8.2).
#
# The forecast is the binary branch of our pre-registered, market-benchmarked
# forecasting method, frozen 2026-06-22 by a single operator. The market mid is the devigged
# Polymarket "Largest Company end of December 2026" market, retrieved 2026-06-22
# (~$3.6M volume). The standings are public market-capitalization aggregates as
# of the same date. The historical leader record is the public market-cap record;
# year-end leader identities are well documented and the year-end caps are
# approximate. scoring_scenarios is recomputed by notebook.ipynb from the frozen
# p_true (0.85) and null (0.72), running top to bottom with no network access.
paper_id: MKT-013
license: cc-by-4.0

datasets:
  - id: standings
    description: |
      The world's most valuable public companies by market capitalization at the
      2026-06-22 freeze, in trillions of US dollars. Source: public
      market-capitalization aggregators (companiesmarketcap.com and reputable
      financial coverage, June 2026). Plotted as Figure 1.
    file: standings.csv
    rows: 5
    schema:
      company: string
      market_cap_usd_tn: float
    source: public-aggregator
    collection_window: [2026-06-22, 2026-06-22]
    license: cc-by-4.0
    sha256: f9aaf3b43005ca942983ced7c4e8ff1c0b9f0c938ef1b2783665a70a83d2f641

  - id: forecast-vs-market
    description: |
      The ATOL forecast (p_true) and the devigged Polymarket market mid per
      candidate for who will be the world's largest company on 2026-12-31, as
      probabilities. The Field row aggregates Amazon, Saudi Aramco, Microsoft and
      Tesla. The market column is the raw Polymarket yes price devigged by the
      overround; the ATOL column is the frozen single-operator forecast. Plotted as
      Figure 2; deviations in Table 1 are derived in the notebook tilt cell.
    file: forecast_vs_market.csv
    rows: 10
    schema:
      candidate: string
      source: string
      p: float
    source: polymarket+forecast
    collection_window: [2026-06-22, 2026-06-22]
    license: cc-by-4.0
    sha256: 13c516a7762908d6277f549af1dad7d9b919a69f6e12b9944ba9594844823a10

  - id: throne-timeline
    description: |
      The identity and approximate year-end market capitalization of the world's
      largest public company, 2011 to 2026 (2026 is the 22 June freeze), in
      trillions of US dollars. Leader identities are from the public
      market-capitalization record (companiesmarketcap.com, 2026; Wikipedia,
      2026) and are well documented; the year-end capitalizations are approximate.
      Plotted as Figure 3; grounds the declared reference-class base rate.
    file: throne_timeline.csv
    rows: 16
    schema:
      year: int
      leader: string
      market_cap_usd_tn: float
    source: public-record
    collection_window: [2011-12-31, 2026-06-22]
    license: cc-by-4.0
    sha256: 994b643841909f709b65d8cc344ba2adc9e0c27817f00d1c22731ef90329ff99

  - id: scoring-scenarios
    description: |
      The Brier score (lower is better) for the ATOL forecaster (p=0.85) and the
      market null (mid=0.72) under each resolution of the claim. Recomputed by
      notebook.ipynb from the frozen p_true and null. Plotted as Figure 5 and
      tabulated in Table 3.
    file: scoring_scenarios.csv
    rows: 4
    schema:
      outcome: string
      party: string
      brier: float
    source: derived
    derived_from: [forecast-vs-market]
    collection_window: [2026-06-22, 2026-06-22]
    license: cc-by-4.0
    sha256: f4125c8d904ed6e29162a2c6e33929f5570bd05c7ab612b86fce4d5f7c81c20d