VERA
How the pipeline works
From a news article to a calibrated, auditable probability — in four phases. The design principle throughout: the language model proposes; deterministic arithmetic and counted history dispose. No single step produces the analysis. The system does.
01 · Ingest & ground

The article is grounded before it is believed.

The trigger article is ingested and its source graded for reliability — a grading that weights every downstream piece of evidence. Actors and countries are resolved against structured registries, and the question the article raises is stamped against a reference class: a counted set of comparable historical cases drawn from the field’s canonical datasets, with a reproducible derivation script and an audit trail.

hard rule — the question and its resolution criteria are fixed here, before any analysis runs
02 · Analyse

Multiple independent readings, structured claims only.

Several independent analytical passes read the article: the surface narrative and what it obscures, the actors and the constraints binding them, the scenario space, the power and economic transmission channels. Each pass emits typed, structured claims — graded counted / sourced / inference — and none of them is allowed to set the headline number.

hard rule — narrative passes cannot touch the probability
03 · Calibrate

The number is arithmetic, not opinion.

The probability starts at the reference class’s counted base rate. Article evidence is weighed in log-odds with source-reliability multipliers, then deliberately dampened against model overconfidence. The scenario partition is reconciled against the calibrated posterior, and where the analyst layer disagrees with the calibrated number, the disagreement is published beside it as a registered position — scored at resolution, never blended in.

hard rule — nothing downstream of calibration can move the number
04 · Adversarial review & publish

The record argues against itself before anyone reads it.

A falsifiability register assembles the case against the analysis — counter-evidence from the counted record, the analytical inversion, dated observables that would prove it wrong. An adversarial peer review scores the run and its warnings are published, not scrubbed. A compliance screen enforces analytical-not-prescriptive language. The record is then written once, append-only, with pre-registered resolution criteria — and scored publicly against outcomes when its horizon arrives.

hard rule — records are never edited after publication; hits and misses score at equal prominence
counted priorsdeterministic arithmetic fail-loud instrumentsappend-only records pre-registered criteriaadversarial review public Brier scoring
The full reasoning for any single analysis — its evidence weights, its register, its peer review — is published on the analysis page itself, in THE FORECAST and THE AUDIT. This page describes the system’s shape; each record carries its own receipts.
VERA NARRATIVE INTELLIGENCE · © 2026 VERA Narrative Intelligence Ltd · VERA analyses are for informational purposes only and do not constitute financial, investment, or legal advice.