Lasso-Selected OHLCV Features for a GLD Threshold Strategy

Summary

This daily GLD strategy applies a fixed linear score to five features derived from price and volume, including short moving averages and interaction terms. The feature weights and normalization constants are embedded in the script, and the strategy holds a long position whenever the score meets a chosen threshold, closing it otherwise. The accompanying description says the features were selected using Lasso and links the implementation to research on temporal feature selection, including polynomial and interaction features.

The page says this implementation corresponds to the study’s best reported Sharpe-ratio result, but it gives no performance figures or backtest settings with which to assess that claim. The model is described as specific to GLD and daily data; applying it to other instruments or timeframes may produce substantially different results. Historical sample choice, execution assumptions, and backtest configuration also affect reported performance. The script includes a commission setting, but the page does not provide further validation or robustness analysis.

Key ideas

• A fixed linear score combines five price and volume features derived from OHLCV data.
• The strategy enters or maintains a long position when the score meets a threshold and closes it otherwise.
• The accompanying description attributes feature selection to Lasso-based temporal feature research.
• The model is intended for GLD on daily data, and transfer to other settings is uncertain.
• The page supplies no performance figures or detailed backtest configuration for independent assessment.

Tags

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