Signal Forensics Framework

The Signal Forensics Framework is the research procedure used by the Event Intelligence Desk to produce a single directional classification on a public equity. It separates headline sentiment from estimated business materiality and treats market reaction as evidence to interpret, not as a basis for direction.

Primary and secondary disclosures are collected across regulatory filings, company communications, trade press, and listed derivatives data. The system evaluates source consistency before any synthesis step. A source set that fails reconciliation is not advanced to classification.

Catalysts are classified by type — guidance revisions, regulatory disclosures, sector rotations, headline dislocations, demand or margin inflections, supply-chain re-ratings. Each class has historical distributions of materiality and timing that inform downstream scoring.

Materiality scores the estimated business impact of the catalyst on the affected company, not the magnitude of the headline. The system separates headline sentiment from materiality explicitly. A score in the upper band indicates structural rather than sentiment-driven impact.

Narrative dislocation measures the divergence between current market narrative and the reconciled source set. The system identifies whether the market reaction appears narrative-led or cash-flow-led. High dislocation with low materiality favors HOLD; high dislocation with high materiality favors a directional read.

BUY, SELL, or HOLD is the final output of the framework, not its entry point. Classification requires convergence of materiality, narrative dislocation, and source confidence. The signal is the output. The reasoning is the edge.

Horizons are selected from a discrete ladder calibrated to the catalyst class and to the liquidity regime. A horizon without a named trigger or window is not admissible.

Every signal is published with a stated invalidation level. Invalidation defines the price, volume, or source condition under which the classification is wrong. A signal without an invalidation level is not published.

The framework produces conditional research, not advice. Source sets can be incomplete. Market conditions change quickly. Historical distributions of catalyst outcomes do not guarantee future results. The framework explicitly does not optimize for a fixed win rate.