Future Data and Backtesting Risks in a Real-Time Stock Screener
Summary
This community discussion raises concerns that a Chinese-language stock selection service may use data or screening conditions that introduce future information into historical tests. Participants report problems with constituents, concepts, and other selection fields, as well as data accuracy and calculation precision. The exchange distinguishes the tool’s intended use for real-time stock selection from the additional challenge of reproducing its behavior in a backtest.
The platform representatives quoted in the discussion say that some functions were removed from the backtesting environment because they could create a gap between simulated and live results, and that fixes were underway. A practical recommendation is to use the screener to explore an idea, then implement the logic independently and validate the data and timing. One commenter suggests adding the current date to a condition as a way to filter many suspected future-data cases, but the discussion does not establish this as a reliable general remedy. It provides no systematic audit or performance evidence, so the specific claims remain anecdotal.
Key ideas
- A real-time stock screener may produce historical selections that rely on information unavailable at the time.
- Differences between live screening and backtest behavior can make simulated results unreliable.
- The discussion reports concerns about future data, data accuracy, and calculation precision.
- It recommends independently implementing and checking screening logic before relying on backtests.
- A suggested date filter is anecdotal and is not demonstrated as a comprehensive solution.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.