Skip to content
All library documents

Band-Pass Filter Crossovers with Rolling RMS Thresholds

Article Strategy library · Author: ChaoZhang

Summary

This strategy generates long and short signals from a band-pass filter, rather than from the Bollinger Band and mean-PB relationship described in the accompanying prose. The code computes a recursive filter from price using fast and slow periods, estimates a rolling root-mean-square envelope, and compares the filter with the envelope boundaries and zero. A rising filter crossing one of those levels produces a long entry; a falling filter crossing downward produces a short entry. The stated defaults are 41 and 54 periods, and the strategy can enter in either direction.

The document reports that backtest returns were satisfactory, but supplies no results, performance measures, or evidence to evaluate that claim. It also warns that historical-data indicators can react poorly to sharp moves, parameter choices affect signals, and macro changes may impair performance. The explanatory text and source code do not fully align: the prose describes Bollinger Bands and the code uses a filter with RMS thresholds. Treat its trading claims cautiously and validate the actual implementation before drawing conclusions.

Key ideas

  • The source code derives a recursive band-pass series from price using fast and slow periods.
  • Rolling RMS values form upper and lower thresholds around the filtered series.
  • Crosses of the thresholds or zero, combined with the filter’s recent direction, trigger long or short entries.
  • The prose describes a Bollinger Band and mean-PB method, while the included code implements a different signal system.

Tags

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