Every top-tier call published in the Forecast Brief since 2017 — reproduced here with the issue in which it appeared, the confidence interval stated at the time, the anticipation gap, and the verified outcome. The ledger is not a marketing artefact. It is the working file of an independent forecasting desk, opened for institutional review.
Document last revised 14 February 2026. The 2025 GCSP audit cycle closes 31 March 2026; figures below carry the 2024 attestation pending the new release.
The annual audit is not a review of opinion. It is a structural examination of each Forecast Brief against the events that subsequently materialised, conducted against a written protocol first agreed in 2019 and revised in 2023.
The audit examines three things. First, the event definition: each forecast is reproduced as a discrete proposition with an explicit trigger condition, and the audit confirms that the trigger was unambiguous at the time of publication and matched, in good faith, the event as it later unfolded. Second, the confidence interval: every call in the Forecast Brief is issued with a stated probability band, and the audit records whether the realised outcome fell inside that band — a hit is recorded when the directional movement and the structural class of event match the published interval. Third, the outcome measurement window: because the publication is forward-looking over an eighteen-month horizon, the audit scores calls on a rolling basis, retiring them to the long-form ledger only after the full window has elapsed without material revision.
Directional accuracy is reported as the share of calls whose realised direction — expansion or contraction of a structural pressure — matched the published direction, irrespective of timing. The headline figure of 87% is calculated against the universe of all calls issued in Issues 73 through 109, the window covered by the 2024 attestation cycle; the 91% figure on top-tier inflection points is drawn from the same window restricted to the twelve highest-consequence calls of each issue.
The audit is conducted under a standing memorandum of understanding between the editorial desk and the GCSP's Risk and Resilience Programme. Methodology notes, the 2019 protocol, and the 2023 revision are available on request to credentialed institutional readers via the subscriptions office in Geneva.
Six reproduced entries from the long-form ledger. Issue number refers to the Forecast Brief in which the call first appeared; confidence interval is reproduced as published; anticipation gap is measured in months between the cover date of the issue and the month the trigger event materialised.
| Event as Forecast | Issue | Published Confidence | Anticipation Gap | Verified Outcome |
|---|---|---|---|---|
| 2024 Taiwan Strait naval standoffStructural test of the post-2022 deterrence equilibrium | Issue 103 · Aug 2023 | 74 – 82% | 9 months | HIT Materially matched within stated band |
| German coalition collapse, November 2024Three-party federal arrangement ends before term | Issue 98 · Feb 2024 | 71 – 79% | 9 months | HIT Materially matched within stated band |
| Argentine peso liberalisationEnd of the crawling-peg regime and one-shot devaluation | Issue 99 · March 2024 | 68 – 76% | 11 months | HIT Materially matched within stated band |
| Sahel state withdrawal from ECOWASCoordinated withdrawal of three member states from the regional bloc | Issue 91 · Sept 2023 | 62 – 71% | 14 months | HIT Materially matched within stated band |
| Red Sea shipping disruptionSustained commercial rerouting lasting more than six months | Issue 95 · Nov 2023 | 78 – 86% | 5 months | HIT Materially matched within stated band |
| India–Gulf state security compactFormal bilateral defence arrangement above joint-exercise level | Issue 100 · April 2024 | 55 – 64% | 16 months | PARTIAL Lower-tier arrangement materialised; full compact did not |
Partial calls are recorded as such and are not counted toward the 91% top-tier figure. The full ledger of all 121 issues, including lower-tier calls and analyst dissents, is available to credentialed readers through the Members' Forecast Brief archive.
Each dossier reproduces the call as it appeared in the Forecast Brief, identifies the analyst who signed it, and records the verified outcome against the audit protocol.
A structural reassessment of commercial routing through the Bab el-Mandeb, anticipating sustained disruption of more than six months and a permanent shift in container insurance premia. The call recorded two dissenting analyst notes in the original issue, both preserved in the archive.
Anticipated the breakdown of the three-party federal arrangement nine months before the formal end of the coalition. Probability band: 71 – 79%.
Projected the abandonment of the multi-band crawling peg and a one-shot devaluation of more than twenty percent. Probability band: 68 – 76%.
Anticipated a three-state coordinated withdrawal from the regional bloc. Probability band: 62 – 71%.
Projected a structural shift in EU member-state procurement patterns for grid-scale storage, independent of the spot price cycle. Probability band: 65 – 74%.
The Forecast Brief is restricted to credentialed institutional readers. The review is conducted by the subscriptions office in Geneva and examines professional standing, institutional affiliation, and intended use of the analysis. Most files are processed within five business days.
