Changing Factor Lists Across Rolling Training and Prediction Windows

Summary

This BigQuant forum answer explains how to use a different feature set in each iteration of a rolling model workflow. The example loop updates training and test start and end dates for each rolling window, disables the backtest module during model runs, and calls the workflow with a parameter dictionary. To vary factors by period, the answer says to update the feature-list module’s parameter inside the loop before each run, using the factor combination assigned to that window.

The guidance is a practical implementation pattern for walk-forward or rolling research, where both the date ranges and model inputs can change between iterations. It does not specify how factor sets should be selected, whether selection is based only on information available at the time, or how to prevent leakage and overfitting. No comparative results or validation procedure are provided, so the answer addresses workflow configuration rather than establishing that changing factors improves predictive performance.

Key ideas

• A rolling workflow can update training and prediction date ranges for each iteration.
• The feature-list parameter must be changed within the loop when each period uses a different factor set.
• The updated parameters should be passed to the workflow run for that rolling window.
• The answer describes configuration mechanics but does not explain how to select factors or prevent look-ahead bias.

Tags

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