GLD Long Strategy Using Lasso-Selected OHLCV Features

Summary

This research implementation turns a sparse linear model into a long-or-flat signal for GLD on the daily timeframe. The model uses five features derived from short moving averages of volume and price relationships among open, high, low, and close. It standardizes those features, combines them with fixed coefficients, and enters or maintains a long position when the resulting score meets a configurable threshold; otherwise, it closes the position. The description says the features were selected with Lasso and relates the approach to research on temporal feature selection and feature interactions.

The published description identifies the implementation as instrument-specific and says the reported study result depends on its historical sample, execution assumptions, and backtest settings. It offers no detailed performance figures in the supplied text, so the model’s effectiveness cannot be assessed from this document alone. Fixed coefficients and threshold may not transfer to other symbols or timeframes, and historical fit does not establish future performance.

Key ideas

• The strategy converts a linear score from OHLCV-derived features into a long-or-flat position.
• Its feature set includes volume, candle-price relationships, and interactions between price measures and volume.
• The score is compared with a configurable threshold to enter or close the long position.
• The implementation is designed specifically for daily GLD data, so transfer to other markets is uncertain.
• The supplied description gives no detailed performance evidence, and results depend on sample and execution assumptions.

Tags

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