backtest-diagnose

To diagnose a failing Python trading backtest, inspect artifacts like metrics.csv and equity.csv to form a baseline, review code files such as signal_engine.py for logic issues, and analyze config.json for parameter mismatches.

Backtest underperformance often stems from logic mismatches in code/signal_engine.py or incorrect settings in config.json. Reading trade artifacts like trades.csv helps identify the root cause by comparing expected and actual signal generation behavior.

The best way to fix a broken backtest is to apply targeted code or configuration adjustments after diagnosis, then rerun the backtest to verify improvements by inspecting updated metrics and equity data for stability.

Yes, backtest diagnosis involves inspecting config.json and Python signal engine files to identify mismatches or logic issues, reading output artifacts to establish a diagnostic baseline, and verifying fixes through reruns.

To verify a backtest fix, rerun the backtest after applying changes and inspect the updated metrics.csv and equity.csv artifacts to ensure the trading logic improvements are stable and resolve the original issue.

Troubleshooting an underperforming backtest requires reading artifacts from the artifacts directory, specifically metrics.csv, equity.csv, and trades.csv, to establish a diagnostic baseline before inspecting source code.