PORTAL / LEDGER-01 / CALIBRATION LEDGER
The Calibration Ledger.
Every dated, resolvable claim we publish is armed at publication and scored at resolution — Brier standard, verdict within one cycle, unresolved-expired takes a flat 0.30 penalty, and every correction carries a process fix. Entries carry the running record's CL- / COR- series (GS-RRA cycle line); the standing orders governing them live in DIV-01. Calibration means 70% claims resolve roughly 70% of the time — not "being right often."
LEDGER-01.1 — Scoring Standard
How entries resolve.
BRIER STANDARD · NO RETROACTIVE ENTRIES · NO MOVED GOALPOSTS
| Rule | Text |
|---|---|
| R-1 | Armed at publication. No claim enters the Ledger retroactively. Every entry is dated, falsifiable, and carries its resolution conditions on the day it publishes. |
| R-2 | Verdict within one cycle. TRUE, FALSE, or PARTIAL (scored as a 0.5 outcome, with the reason stated) — entered within one reporting cycle of resolution. |
| R-3 | Unresolved-expired takes a penalty. A flat 0.30. Indefinite deferral is banned — a prediction we stop tracking still counts against us. |
| R-4 | Corrections in the same ink. Every correction publishes in the same record as the wins, within two cycles, each with a process fix so the same failure cannot repeat silently. |
LEDGER-01.2 — Live Record
Notable entries,
misses included.
SELECTED FROM THE ACTIVE RECORD · FULL EXPORT WITH EACH MONTHLY CALIBRATION REPORT
| Entry | Claim | Resolution mechanism | Status |
|---|---|---|---|
| CL-041 | Islamabad Memorandum window | MOU window assessment; autopsy published at closure — sabotage traced to internal hardliners on both sides | False |
| CL-054R | Chabahar exclusion zone — coordinates in the indexed record | RESOLVED-FALSE 25 Sep (SW-2026-005, Brier 0.3025, signed FINAL): coordinates absent from every indexed family since 10 Sep — 15 days at execution; NAVAREA IX / IMO pulls clean. Edge-test follow-on CL-080 holds armed | False |
| CL-056 | U.S. warplane-damage disclosure | Service-level candor confirmed; presidential/Pentagon-level non-disclosure — process fix: disclosure claims must name the podium (SRR-001, Brier 0.0100) | Partial |
| CL-057 | Houthi Bab el-Mandeb non-disruption pledge | 30-day behavioral test against pledge terms — holding, day 14 of 30 (WPS № 1, 25 Sep) | Active |
| CL-059 | Saudi insured-autonomy | Graduated behavior → policy progression tracked against stated thresholds | Active |
| CL-060 | Iranian toll flexibility | 10% freight-charge suspension held through the window (SRR-001, Brier 0.1600) | True |
| CL-065 | Patriot interceptor inventory (~65% expended) | Confirmed in budget language — $22B supplemental cleared markup with an inventory-transparency amendment | Active |
| CL-067 | Salalah rescheduling | Rescheduled for 21 Sep inside the 14-day window; dormancy tripwire disarmed | Superseded |
| CL-068 | Qeshm cargo-ship strike attribution | Mine signature confirmed, no claimant; entry lacked a probability at issue — UNSCORED, and a binding rule adopted: no attribution entry without probability + resolution date | Unscored |
| CL-074 | Salalah 21 Sep round will produce a documented deliverable (p 55%) | RESOLVED-TRUE on the pre-registered literal reading (agenda verified; Brier 0.2025) — the letter held though the room did not convene. The anti-goalpost rule's cost, carried openly. | True |
| CL-075 | Chabahar "maps published 17 Sep" claim | No indexed confirmation ever existed outside our own briefs — worst miss to date (Brier 0.6400); fired the drafting-process review now binding on all drafting | False |
| CL-076 | Saudi acceptance of the 21 Sep Salalah round | Two source families, two rows → CONFIRMED (Brier 0.0625) | True |
| CL-081 | German state elections: AfD plurality without majority in BOTH Berlin and Mecklenburg-Vorpommern (p 55%) | Conjunct failed on the Berlin leg (AfD third, ~16.0%); MeckPomm leg met (~38%, firewall held) — INTERIM FALSE pending certified results; Brier 0.3025 pre-staged. A priced miss, logged in the same ink as the wins. | Interim False |
| CL-084 | Salalah diplomatic round — dated reconvene announcement by 30 Sep (p 50%) | RESOLVED-FALSE 30 Sep (Brier 0.2500): "no new date set" universal across every indexed family; Omani FM primary record silent since 13 Sep. Spirit noted: the mediator channel works at message level — it never produced a date. | False |
| CL-085 | A second indexed family confirms the Salalah postponement was at Saudi Arabia's request by 29 Sep (p 40%) | RESOLVED-FALSE 29 Sep (Brier 0.1600): the verbatim request attribution stayed single-origin Iranian (Baghaei/IRNA) through 15 days of all-family screens. Spirit noted: the Saudi role is real and two-family Western; the request attribution is not. | False |
| CL-086 | US–Iran track — dated leader-level meeting announcement by 30 Sep (p 30%) | RESOLVED-FALSE 30 Sep (Brier 0.0900): the Qatar relay landed at mediator level — the US response reached Pezeshkian's Cabinet table 30 Sep, the dispute narrowed to sequencing — and still no dated leader-level announcement exists. Motion real; the letter's level unmet. | False |
| CL-DEF-1 | Self-audit: OTH-gate deferral — CLOSED by execution | OTH gate executed 28 Sep — NOT CERTIFIED (OTH-CRT-001); tiers demoted to unproven signal; soft-landing §8 executed (non-commutative dynamics retired, spectral demoted, PGE retired); re-run path of record — read in full | Not certified |
LEDGER-01.3 — Corrections Register
Corrections,
in the same ink.
PERMANENT · NUMBERED · PUBLIC · EACH WITH ITS PROCESS FIX
| Ref | Finding | Process fix |
|---|---|---|
| COR-1 | Syria integration assessment overstated; dependent entry repriced (CL-009, 60% → 55%) | Demonstrative-event claims are auto-quarantined at drafting |
| COR-2 | Iranian tiered toll architecture (PGSA authority, "Hormuz Safe" insurance) operational since May but missed by the brief line | Rent-system baseline added to all corridor assessments |
| COR-3 | Single-day flow snapshots used in headlines overstated day-to-day volatility | Headlines carry 7-day rolling averages only; single-day snapshots banned |
| COR-4 | Scheduled events referenced without independent verification | Mandatory verification of any scheduled event within 48 hours of reference |
| COR-5 | DRC Ebola entry carried a false premise — a PHEIC "declared 17 Sep" never existed; the WHO primary record shows the determination of 17 May 2026 | Primary-record verification before entry; affected section restated; dependent entry rebaselined to WHO figures |
| COR-6 | A market figure (~$114 Brent) printed without a dated source; the week closed near $103–104 | Dated-print rule now binding — every market figure carries a dated source or does not print; spot-checks in Sunday audits |
| COR-7 | Two briefs narrated a convened Salalah session that never occurred — the postponement sat in the wire record two days unswept; session-output inventory empty across every family. Highest-severity correction to date. | F-6 cancellation check binding: no scheduled event narrated as occurred without a 72-hour postponement sweep and two-row occurrence confirmation — occurrence is a claim like any other |
| COR-8 | A printed War Powers line conflated count, date, and sponsor — "Jayapal / 214–208 / 16 Sep" — with the 23 Jul H.Con.Res.89 measure; the primary record is Clerk Roll Call 307: H.Con.Res.93, 220–204, 15 Sep, Moulton | F-8 primary-record rule binding: legislative action prints only from the Clerk roll call / congress.gov; sibling measures must carry measure numbers |
| COR-9 | A sweep printed the 24 Sep Senate vote under the wrong measure identity (H.Con.Res.93 vs. the actual H.Con.Res.89 — count, date, and vote number all correct); the erroneous print stood ~15 hours. Committed two days after F-8 entered binding law — the rule's text was sufficient; its execution at the measure-number step was not | The conflation check now attaches to the sweep template itself: any legislative print names its measure, its chamber, its roll-call number, and the page the three were read from — four fields, one line, every time |
| COR-10 | A signed brief's ledger line divided a two-resolution day by 19 instead of 20 — Brier printed 0.1834 over 19 scored; the true figure is 0.1742 over 20, with the FALSE tally one short. Caught by the publication desk's pre-deploy reconciliation: no public page ever carried the error, and the correction improves the number it fixes | Combined staged arithmetic now prints its base-and-increment line ("base 18, +2 resolutions → ÷20") — verification sees the increment, not just the quotient |
> early-warning tier system: TESTED 28 SEP 2026 — DID NOT CLEAR THE GATE · demoted to unproven signal
> validation gate: executed 28 SEP 2026 — verdict NOT CERTIFIED, published in full as committed (OTH-CRT-001)
> soft-landing: executed of record — non-commutative dynamics retired, spectral demoted, rebuild classical. The standard held; the model did not.
> validation gate: executed 28 SEP 2026 — verdict NOT CERTIFIED, published in full as committed (OTH-CRT-001)
> soft-landing: executed of record — non-commutative dynamics retired, spectral demoted, rebuild classical. The standard held; the model did not.