Diagnosing Inconsistent Option Strategy Backtests
Summary
A short forum exchange raises a reproducibility issue: the same options strategy reportedly produced different returns in two backtests. The original poster wonders whether running several strategies at the same time caused the discrepancy. A respondent recommends starting with the trade records and checking whether a fresh backtesting engine instance was created.
The exchange offers a practical first-pass diagnostic rather than a demonstrated explanation. Comparing fills can reveal whether the runs differ in execution records, while checking engine initialization may help uncover reused or inconsistent state. However, the thread does not show the trade logs, describe the backtesting framework, establish that concurrent runs were responsible, or report a confirmed fix. It therefore serves as a troubleshooting prompt, not evidence that simultaneous strategy testing necessarily changes results. Further investigation would need to control the data, settings, engine lifecycle, and execution assumptions across runs.
Key ideas
- Different returns from repeated tests of the same options strategy warrant checking the recorded trades.
- Confirm whether each run uses a newly initialized backtesting engine.
- The thread does not establish that concurrent strategy runs caused the discrepancy.
- Reproducibility checks require comparing data, settings, engine state, and fills.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.