signal-postmortem

Postmortem analysis for trading signals aggregates post-trade records to compute accuracy, false positive rates, and regime correlations. It classifies outcomes into true positives, false positives, missed opportunities, and regime mismatches to validate decision quality.

Generate weight adjustments by processing signal outcomes over 5-day and 20-day holding periods within a rolling 90-day window. The computed results apply minimum sample thresholds to produce weight feedback JSON files suitable for downstream signal-aggregation systems.

Creating a signal improvement backlog involves quantifying skill performance from postmortem records and identifying systematic errors. This Skill outputs backlog YAML entries to drive skill improvement based on historical false positives and regime mismatches.

Postmortem records measure signal accuracy by computing regime correlations alongside standard performance metrics. By evaluating outcomes across distinct holding periods, you can identify specific market conditions causing signal degradation or mismatches.

Validating post-trade signal outcomes requires enforcing minimum sample thresholds within rolling 90-day windows. This ensures statistical reliability before generating summary statistics, weight updates, or backlog items for downstream operational workflows.

Regime mismatch classifications occur when postmortem analysis identifies systematic errors caused by specific market conditions. Computing regime correlations helps surface these mismatches, generating feedback to adjust signal generation logic and aggregator weights.