Why Backtests and Simulations Can Produce Different Results
Summary
A trader reports that the same strategy produces different results across backtest runs and simulated trading, including differences in selected securities, order quantities, and execution prices. The discussion offers possible explanations: randomness or use of future information in the strategy, and differences in order processing.
A later reply identifies order submission functions that allocate a percentage of the portfolio as a source of small holding differences when orders are processed in a different sequence. The exchange is a brief troubleshooting discussion rather than a systematic investigation: no code or detailed reproduction is included, and the early claims about future data are not confirmed. It nevertheless highlights that execution order and platform behavior can affect portfolio weights and make backtest and simulation results diverge, even when the intended strategy is unchanged.
Key ideas
- Backtest and simulated trading results can differ in security selection, order size, and execution price.
- Randomness and future data are suggested as possible causes, but the discussion does not verify them.
- Order sequence can change holdings when orders allocate a percentage of portfolio value.
- The exchange provides a troubleshooting lead but lacks code and a reproducible diagnosis.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.