Backtest Diagnosis
Overview
Use this skill when a user reports that a backtest failed, raised an error, or produced poor results.
Diagnostic Workflow
- Read existing artifacts: use
read_fileto inspectartifacts/metrics.csv,equity.csv, andtrades.csv - Read the code: use
read_fileto inspectcode/signal_engine.pyandconfig.json - Classify the issue: determine the root cause using the error taxonomy below
- Apply the fix: use
edit_fileto modify the code, then rerun the backtest - Verify the fix: use
read_fileto inspect the newmetrics.csv
Error Taxonomy
Runtime Errors (exit_code != 0)
| Error Type | Common Cause | Fix |
|---|---|---|
| ImportError | Missing dependency | bash("pip install xxx") |
| KeyError | DataFrame column-name mismatch | Check the actual column names in data_map |
| IndexError | Empty data or insufficient length | Add length checks |
| TypeError | Incorrect signal type | Ensure the return value is pd.Series |
Logic Bugs (Backtest Succeeds but Results Are Abnormal)
- Zero trades (
trade_count=0): signal-logic bug. Conditions are too strict, so the signal st…
