Execution Gaps Between Backtests and Simulated Trading

Summary

This forum post raises operational problems encountered after submitting previously successful stock strategies for simulated trading. The author reports that some buy instructions appear a day early and are later marked invalid, that incomplete purchases can continue over subsequent days and disrupt an intended daily rebalance, and that simulated results can differ substantially from backtests. The examples mentioned include a machine-learning stock selection strategy for Beijing Stock Exchange listings and a multifactor strategy with index timing.

The post asks whether an order can be treated as complete once most of its intended allocation has been bought, allowing a small unfilled balance to remain idle rather than generating repeated small purchases. This highlights how order handling, timing, and residual cash can change realized holding periods and strategy behavior. It does not provide answers, diagnostic evidence, or a proposed general fix, so the issues remain questions for the platform or strategy implementation. The observations are user reports rather than controlled comparisons, and no measured performance impact is supplied.

Key ideas

• The author reports that simulated orders can be rejected after signals are generated ahead of the intended trading date.
• Partial fills may lead to small follow-up purchases that extend a strategy’s effective rebalance period.
• The post asks about considering an allocation substantially filled once a high fraction has been purchased.
• It reports differences between backtest and simulation results without isolating their causes.
• The document presents questions and user observations, not a validated diagnosis or solution.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.