Adding Index Data to Rolling-Training Backtest Preparation
Summary
The discussion addresses missing broad-market risk-control data in a rolling-training workflow. Its suggested remedy is to load benchmark index history directly in the backtest data-preparation function instead of relying on an external data merge. The example reads index closing prices with extra lookback history and derives a one-day index return, which can supply the benchmark series needed by the workflow.
The response also says to adjust rolling-training parameters to fit the use case and to pass the identifier for the module connected to the trade stage as the prediction identifier. A trade module is not mandatory; the workflow may connect to other modules. This is a short implementation-oriented answer rather than a tested explanation: it does not establish why the original risk-control data was missing, specify all required parameters, or demonstrate that the suggested setup fixes every configuration.
Key ideas
- Load benchmark index data inside the backtest data-preparation function when external merging causes missing risk-control data.
- Use historical index closes to calculate a daily benchmark return.
- Set the prediction identifier to the module connected to the trading stage.
- Rolling-training workflows can connect to modules other than the trade module.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.