Signal Forensics Framework

The Event Intelligence Desk produces conditional research on discrete, dated events affecting listed equities. The unit of analysis is the event rather than the security: coverage is initiated by a catalyst and terminates at resolution or invalidation, and is not maintained continuously thereafter. Each published position is a conditional statement carrying an explicit horizon and an explicit invalidation condition.

The desk operates exclusively within the public information set. It does not solicit, receive, or act upon material non-public information, and it does not rely on expert networks, private channel checks, or any source whose value derives from subscriber access. The analytical claim is therefore not informational advantage but interpretive discipline: that a public record read consistently and in full yields conclusions the prevailing account of that record does not.

Positions are constructed from primary disclosure — regulatory filings, company communications, and listed derivatives data — read against secondary coverage rather than through it. Sources are weighted by proximity to the issuer and by verifiability. Inconsistency between sources is treated as a finding in its own right, not as noise to be averaged away. A source set that fails reconciliation is not advanced, and where the primary record and the prevailing account of it diverge, the divergence is documented and preserved rather than resolved by assumption.

Materiality is the estimated effect of the catalyst on the issuer's economics. It is held strictly distinct from headline magnitude, from price reaction, and from sentiment, and none of the three is admitted as a proxy for it. Market reaction is an object of interpretation rather than evidence of direction. A large move is information about positioning and expectations; it is not information about business impact.

The framework that produces a classification is not published. No decision rules, weightings, thresholds, catalyst taxonomy, or calibration procedure appear on any surface. This is a deliberate and permanent limit rather than documentation pending release. Research procedure is the firm's working asset, and any description detailed enough to be useful to a reader is by construction detailed enough to be reproduced. Verification is offered instead through the published record: an ex ante, timestamped, and falsifiable account of the framework's output, which is the more demanding of the two standards and the only one that constitutes evidence.

Every output is conditional. A position states the conditions under which it holds, the window over which it is expected to resolve, and the observable conditions that would render it wrong. A position that cannot be specified in those terms is inadmissible and is not published. Admissibility is a precondition rather than a quality score; where a catalyst does not permit a specifiable position, the desk declines to publish one.

A classification is BUY, SELL, or HOLD. It is the terminal output of the assessment rather than its point of departure, and it is directed at no particular reader. HOLD is a substantive conclusion rather than the absence of one: the finding that an observed reaction is adequately explained by the reconciled record. No classification is published without an accompanying horizon and invalidation condition.

Each position specifies the window over which it is expected to resolve and the observable event that would constitute resolution. Horizons are stated, never implied. A horizon without a named trigger, or a trigger without a bounded window, is inadmissible. An unbounded thesis cannot be shown to have failed, and a claim that cannot fail is not research.

Every position carries a stated invalidation condition: the price, volume, or disclosure event under which the classification is to be treated as wrong. Invalidation is specified ex ante and is not revised afterward. It is recorded when it occurs, and a met invalidation closes the position irrespective of subsequent price action. A position without an invalidation condition is not published.

Closed classifications are published with their original date, horizon, and invalidation condition, alongside the manner of their resolution. Entries are not revised, withdrawn, or reordered after publication, and invalidated positions remain in the record on the same terms as resolved ones. The record is the firm's account of its own output and the basis on which it asks to be assessed. It is presented in full rather than in selection, because a record shown selectively is not a record.

The desk produces conditional research of general and impersonal circulation, not personalized advice. The public information set is sometimes incomplete, and reconciliation cannot recover what has not been disclosed. Interpretation of a correctly assembled record can still be wrong. Regimes change faster than catalysts resolve, and a position correct on an issuer's economics can be invalidated by conditions unrelated to them. Historical outcomes carry no implication for future ones. The desk does not optimize for a fixed win rate, and would treat a stable one as evidence of selection rather than of method.