The Calibration Ledger
16 predictions sealed with locked probabilities, resolution dates and falsifiability criteria. A sealed probability is never rewritten.
CAL-01 — US federal debt held by the public exceeds 105% of GDP in an official CBO or FRED reading.
Sealed at 78% on 2026-07-03. Resolves by 2029-12-31. Status PENDING.
Correct if: CORRECT if FRED series FYGFGDQ188S (federal debt held by the public, % of GDP) prints ≥105 for any quarter, or CBO reports an actual — not projected — fiscal-year figure ≥105%, on or before 2029-12-31. Baseline at sealing: ≈100% of GDP (CBO, 2025); CBO baseline crosses the 1946 record (~106%) in 2029.
CAL-02 — No direct kinetic exchange between US and Chinese armed forces — no incident in which the two militaries exchange live fire, in any theater, at any scale.
Sealed at 82% on 2026-07-03. Resolves by 2029-12-31. Status PENDING.
Correct if: CORRECT if, through 2029-12-31, no exchange of live fire between US and PRC military forces is confirmed by either government or independently documented by three or more major international news organizations. Collisions, lasing incidents, cyber operations and proxy engagements do not count — live fire only.
CAL-03 — More than 60% of US adults report using generative-AI tools at least weekly in a major national survey.
Sealed at 62% on 2026-07-03. Resolves by 2029-06-30. Status PENDING.
Correct if: CORRECT if Pew Research or an equivalent probability-sampled national survey (Gallup, AP-NORC, Reuters Institute) published by 2029-06-30 finds >60% of US adults using generative-AI tools weekly or more often. Baseline at sealing: Pew (June 2025) found 34% of US adults had ever used ChatGPT.
CAL-04 — The S&P 500 avoids a drawdown of 25% or more between sealing and mid-2027.
Sealed at 78% on 2026-07-03. Resolves by 2027-06-30. Status PENDING.
Correct if: CORRECT if, before 2027-06-30, the S&P 500 daily close never falls 25% or more below its highest daily close recorded after 2026-07-03. One clean threshold, daily closing data only, no interpretation. This entry resolves at the first public scoring in July 2027.
CAL-05 — A formal ceasefire or peace agreement in the Russia–Ukraine war holds for at least six consecutive months.
Sealed at 55% on 2026-07-03. Resolves by 2028-12-31. Status PENDING.
Correct if: CORRECT if a signed ceasefire, armistice or peace agreement is followed by six consecutive months in which ACLED-recorded conflict events on Ukrainian territory fall at least 80% below the 2025 monthly average, with that six-month window completed on or before 2028-12-31.
CAL-06 — The US dollar falls below 45% of allocated global FX reserves in IMF COFER data.
Sealed at 8% on 2026-07-03. Resolves by 2030-12-31. Status PENDING.
Correct if: CORRECT if any quarterly IMF COFER release through 2030-12-31 shows the US dollar share of allocated reserves below 45%. Baseline at sealing: ≈57–58% (COFER, 2025). Sealed LOW deliberately — de-dollarization is real, but reserve inertia is measured in decades, not news cycles. The base case says NO.
CAL-07 — At least one calendar year from 2026 through 2029 sets a new global temperature record.
Sealed at 76% on 2026-07-03. Resolves by 2030-06-30. Status PENDING.
Correct if: CORRECT if NASA GISTEMP or NOAA NCEI declares any of 2026, 2027, 2028 or 2029 the warmest calendar year on record in their annual analyses, published by 2030-06-30. Record holder at sealing: 2024, ≈1.55°C above pre-industrial (WMO) — the first calendar year above 1.5°C.
CAL-08 — Bitcoin records a daily close above $150,000 on a major reference rate.
Sealed at 58% on 2026-07-03. Resolves by 2028-12-31. Status PENDING.
Correct if: CORRECT if BTC-USD posts a daily close above $150,000 on Coinbase or the CME CF Bitcoin Reference Rate at any point between sealing and 2028-12-31. Intraday wicks do not count; a single qualifying daily close resolves the entry.
CAL-09 — NASA lands astronauts on the lunar surface (Artemis III) by the end of 2028.
Sealed at 25% on 2026-07-03. Resolves by 2028-12-31. Status PENDING.
Correct if: CORRECT if a crewed Artemis III lunar surface landing is completed and confirmed by NASA on or before 2028-12-31. Sealed LOW deliberately: the official schedule says mid-2027, but Starship HLS milestones and suit readiness have slipped before, and GAO has repeatedly flagged the timeline. The base case says NO.
CAL-10 — Global solar installations exceed 900 GW(dc) in a single calendar year by 2030.
Sealed at 66% on 2026-07-03. Resolves by 2031-06-30. Status PENDING.
Correct if: CORRECT if IEA, BloombergNEF or SolarPower Europe reports global annual PV additions above 900 GWdc for any calendar year through 2030, with final figures published by 2031-06-30. Baseline at sealing: ≈599 GWdc installed worldwide in 2024 (BloombergNEF).
CAL-11 — A formal Russia–Ukraine ceasefire, armistice or peace agreement is signed and still in force on 2027-12-31.
Sealed at 48% on 2026-07-11. Resolves by 2027-12-31. Status PENDING.
Correct if: CORRECT if on 2027-12-31 a signed ceasefire, armistice or peace agreement between Russia and Ukraine is in force — signed at any earlier date and not collapsed back into open hostilities — as documented by major international news organizations and ACLED conflict-event data. A collapsed agreement scores INCORRECT.
CAL-12 — 2026 ranks among the two warmest calendar years ever measured, in both major global temperature analyses.
Sealed at 66% on 2026-07-11. Resolves by 2027-02-28. Status PENDING.
Correct if: CORRECT if the annual 2026 analyses from BOTH NASA GISTEMP and Copernicus ERA5, published January–February 2027, rank 2026 first or second among all calendar years on record. Either analysis ranking it third or lower scores INCORRECT. This is the fastest-resolving seal on the ledger.
CAL-13 — A frontier AI lab documents an agent autonomously completing a full week (≥40 hours) of economically-valuable professional work, or METR’s 50%-reliability time horizon reaches 40 hours.
Sealed at 75% on 2026-07-11. Resolves by 2027-12-31. Status PENDING.
Correct if: CORRECT if by 2027-12-31 a METR-published 50%-reliability time horizon reaches 40 hours for any frontier model, OR a frontier lab (OpenAI, Anthropic, Google DeepMind, Meta, xAI) publishes a documented case of an agent autonomously completing ≥40 hours of contiguous economically-valuable professional work, or expert-parity performance across the full GDPval suite.
CAL-14 — Humanoid robots cross the industrial threshold: at least 50,000 units shipped worldwide in calendar 2026.
Sealed at 55% on 2026-07-11. Resolves by 2027-06-30. Status PENDING.
Correct if: CORRECT if full-year 2026 global humanoid shipment tallies published by IDC, TrendForce or Omdia by 2027-06-30 reach 50,000 units or more. Announced orders and backlogs do not count — shipped units only.
CAL-15 — At least one G7 central bank cuts its policy rate by a cumulative 100 basis points or more between July 2026 and end-2027.
Sealed at 38% on 2026-07-11. Resolves by 2027-12-31. Status PENDING.
Correct if: CORRECT if between 2026-07-01 and 2027-12-31 any G7 central bank (Federal Reserve, ECB, Bank of England, Bank of Japan, Bank of Canada) reduces its primary policy rate by ≥100bp cumulative from its July 2026 level, per official policy announcements. A full-point cycle signals recession-grade easing — this entry is sealed as the minority case.
CAL-16 — No uncrewed Starship HLS soft-lands on the Moon before 2028 — the Artemis III critical path slips past 2027.
Sealed at 68% on 2026-07-11. Resolves by 2027-12-31. Status PENDING.
Correct if: CORRECT if no uncrewed Starship HLS completes a confirmed soft landing on the lunar surface by 2027-12-31, per NASA and SpaceX mission confirmations. A confirmed soft landing on or before that date scores INCORRECT. Sealed as the base case given the demonstrated launch cadence and unproven orbital-refueling chain.