What We Publish.
What We Do Not.

An intelligence analysis desk, not a magazine

The unit is constituted as an intelligence analysis desk. It operates editorially in Italy under the FP3NEWS commercial brand of the editorial title FreePress3.News, registered with the Tribunale di Verona · Registro Stampa N. 2149 · R.G.V.G. 7274/2020 · 21 December 2020. Publisher: MODENART SRLS (CCIAA Verona · P.IVA 04522520230 · REA VR-427683). Editor-in-Chief: Iacopo Mutascio (Ordine dei Giornalisti del Veneto). Its output discipline is that of a private analytic unit producing scenario-based intelligence assessments for institutional readers, not that of a generalist news outlet producing daily news flow.

This distinction is not cosmetic. It governs what we accept as evidence, how we commit to forecasts, who we distribute to, and what we expect from our readers. A reader who treats a PULSE INSIDER paper as a news article will mis-read it. A reader who treats it as a probabilistic assessment with declared confidence and committed resolution criterion will read it correctly.

Anti-overclaim discipline

Every analytical claim produced by the unit passes through the §17 operator, the signature mechanism of the riscontro protocol. The §17 operator forces three explicit declarations on every claim:

1. Evidence status — verifiable data, or stated inference. No hedge.
2. Confidence band — explicit percentage interval. No "high confidence" without a number.
3. Resolution criterion — a binarisable event and a date by which the claim will be true or false.

A claim that cannot pass §17 is rewritten or withdrawn. This is the most important single mechanism of the unit's analytical discipline. It is the difference between a calibrated forecast and a confident-sounding opinion.

Why we do not publish analyst identities

The unit operates through the FP3 method — a structured, method-governed and evidence-led analytical method. the FP3 method is the unit's operational modality: method-governed and evidence-led, §17-governed. Cluster mapping and node functions are publicly disclosed at . The framework is a method, not a staff chart — governance-based and editorially accountable. FP3 is a method-governed, evidence-led and accountable unit: the analytical work is led and signed by the Director, augmented by AI devices across the. FP3 publicly discloses its governance, evidence standards, methodological controls and verification criteria. Operational implementation details remain reserved as protected know-how. For the toolchain layer of the architecture, see Article 7 · Tool & Augmentation Disclosure.

This is the standard practice of many private analytic units producing intelligence-grade material for institutional readers. The classical academic publication model — where every contributor is named, with CV and biographical profile — serves a different epistemic function (peer-review traceability for reproducibility). The intelligence analytic model serves a different function: protect contributor identity to preserve the unit's operational integrity, while exposing the analytical product itself to maximum accountability through declared confidence, registered resolution, and ex-post riscontro verification.

There are well-documented institutional precedents for this model: certain Oxford Analytica briefing units, Stratfor private intelligence reports, RUSI restricted publications, and private boutique advisory units typically operate under disclosed-by-product / undisclosed-by-author protocols. The unit follows this established practice.

Who receives what

The unit operates a tiered distribution architecture aligned with the analytical sensitivity of each output.

Routing discipline is enforced under source integrity H3 (Distribution OPSEC). Cross-sector contamination is structurally prevented — for instance, investigative dossiers on Italian media systems are never distributed to wine industry contacts, regardless of subscription tier.

Negative declaration

A doctrine is also defined by what it refuses. The unit does not:

• Produce daily news flow. We publish papers, not articles.
• Issue forecasts without declared probability and registered resolution criterion. Confidence-sounding language without numbers is forbidden under §17.
• Retract failed forecasts. Failed forecasts are scored, kept on file, and discussed in subsequent papers as drift evidence.
• Accept commissioned analysis where the conclusion is implicitly specified by the commissioning party. We work under independence terms only.
• Publish analyst-level CVs, biographies, or contributor identities. See Article 3.
• Operate behind political alignment. The unit carries no party affiliation, no campaign deliverables, no advocacy-cycle output.
• Distribute restricted material to readers who decline to sign an NDA appropriate to the tier. The signature is the protection of the reader as well as the unit.

How to assess our work

A serious institutional reader assessing the unit should run the following three tests:

1. Riscontri test — pull the numbered Archive, identify papers whose resolution criterion has expired, and verify that the riscontros published are consistent with the declared probabilities at the time of publication. If the riscontros converge on or below the case baseline over a sustained sequence, the unit is calibrated.
2. Independence test — identify papers whose conclusions cut against the commercial or political interests of any party the unit is institutionally connected to. If such papers exist and were distributed despite the friction, the unit is operationally independent.
3. Boundary test — request methodology disclosure for a specific Protocol under NDA. The disclosure provided should be specific enough to allow methodology review by a qualified institutional reviewer, while preserving operational reserve. If both conditions are satisfied, the Disclosure Boundary is honest.

These three tests are the right way to assess us. We invite them.

Governance, verification and reserved infrastructure

FP3 publicly discloses the elements required to evaluate its work: source provenance, evidence standards, separation between fact and inference, analytical method, dated forecasts, resolution criteria, corrections, governance and the public verification record. Internal operational infrastructure, configurations and implementation procedures remain reserved professional know-how.

Accountability is at output level. Published claims remain inspectable and falsifiable through their sources, dates, criteria and outcomes. Editorial responsibility remains identifiable under the applicable rules governing the publication.

How weak signals are gated before they reach a forecast

Open-source and social chatter is the fastest, dirtiest vector: high coverage, high speed, low Admiralty grade. The desk does not trust it — it processes it. A signal passes five stages before it goes anywhere near a published forecast: ① Recon (≥1h daily directed scan by the Director) · ② Triage (clustering, noise stripped) · ③ Confirm gate (the hard wall) · ④ Corroborate (mapped against existing theses — confirm, weaken, complicate?) · ⑤ Feed & log (Bayesian update where material, archived with falsifiers).

The contamination gates (stage ③): G1 provenance — walk the claim back to source-zero or it stays a rumour · G2 citation & figure — every number/date/quote checked against a Tier-1 source · G3 synthetic-media — reverse-image & manipulation screening, with the honest caveat that detector false-negatives are rising, so absence of a flag is not proof of authenticity · G4 cross-source triangulation — a single source never promotes a claim above Tier-3.

What it is: a detection layer (where to look, not what is true) · a tripwire for weak signals · a corroboration engine. What it is not: a source of record (social never anchors a claim alone) · a debunking column · a mass-collection operation · an auto-publishing bot — nothing ships without a human signature. The grammar is probabilistic, not adjudicative: an honest miss on the record strengthens the desk more than a claim of infallibility ever could.