signal-postmortem

To analyze trading signal outcomes, you classify closed signals into true positives, false positives, missed opportunities, and regime mismatches, then feed those results back to your edge-signal-aggregator weights and a skill-improvement backlog.

The best way to classify false positives and missed opportunities is by running a postmortem recorder on closed JSON signal records, which categorizes outcomes and generates weight feedback alongside backlog YAML entries.

Yes, you need Python 3.9 or higher and JSON-formatted signal records to run a signal postmortem analysis, with optional API keys required if you need to fetch external price data for the evaluation.

Weight feedback from signal postmortems adjusts risk regime models by taking classified outcome data from 5-day and 20-day holding periods and outputting structured feedback files that recalibrate downstream edge-signal-aggregator weights.

Yes, you can use screener-derived signals with a 5-day holding period for postmortem analysis, as the tool explicitly supports evaluating signals from screeners, the edge pipeline, and other skills over 5-day and 20-day timeframes.

Recording trading signal postmortems generates postmortem JSON files for outcome tracking, weight feedback files for aggregator recalibration, backlog YAML entries for skill improvement, and summary reports for monitoring signal quality.