backtest-diagnose

Diagnosing a backtest with zero trades requires inspecting artifacts like metrics.csv and trades.csv alongside signal_engine.py code to locate offending signal logic or incorrect data assumptions. This structured workflow classifies root causes before applying targeted fixes.

Finding a backtest failure root cause involves stepping through artifact reading, examining signal_engine.py code, classifying issues by error taxonomy, and verifying fixes through rerun validation. This structured workflow ensures precise corrective actions.

Fixing backtest runtime errors requires inspecting available artifacts like config.json and equity.csv, then reviewing signal_engine.py code to differentiate runtime, logic, and data errors. This taxonomy-based classification guides appropriate corrective actions.

You can validate backtest fixes by rerunning the strategy after editing signal_engine.py code. The diagnostic workflow includes validation rules and hard gates confirming metrics, trades, equity data, and exit codes meet quality requirements post-edit.

Underperforming backtests without runtime errors often hide logic bugs or insufficient trade activity. Diagnosing these issues requires reviewing metrics.csv and trades.csv artifacts to identify suspect results and classify underlying logic or data errors.

When diagnosing backtest failures, note the workflow requires available artifacts and code for review including metrics.csv, equity.csv, trades.csv, config.json, and signal_engine.py. It also focuses on ignoring provider-side limitations during error taxonomy classification.