High Pass Filter

Parameters: period = 48 (5–200)

Overview

The High Pass Filter is a digital signal processing technique adapted for financial markets by John Ehlers. It removes low-frequency (long-term trend) components from price data while preserving high-frequency (short-term) components. This makes it particularly useful for identifying cycles, oscillations, and rapid price movements that might be obscured by dominant trends.

The filter acts as the opposite of a moving average. While moving averages smooth out short-term fluctuations to reveal trends, the High Pass Filter removes trends to reveal short-term fluctuations and cycles. This detrended output helps traders identify market cycles and potential turning points with minimal lag.

Interpretation & Trading Signals

Signal Characteristics:

  • Zero Line Crossings: Indicate potential cycle turns or momentum shifts
  • Peaks and Troughs: Mark cycle highs and lows in detrended data
  • Amplitude: Larger amplitudes suggest stronger cyclic movements
  • Frequency: Distance between peaks indicates dominant cycle length

Trading Signals:

  • Buy Signal: Filter crosses above zero after forming a trough
  • Sell Signal: Filter crosses below zero after forming a peak
  • Trend Direction: Above zero = upward momentum, Below zero = downward momentum
  • Cycle Confirmation: Regular oscillations indicate tradeable cycles present

Parameter Guidelines:

  • Short Period (5-20): Captures very short-term cycles, more noise
  • Medium Period (20-50): Balances cycle detection with noise reduction
  • Long Period (50-200): Focuses on longer cycles, smoother output
  • Default (48): Suitable for identifying dominant market cycles

Example Usage

Code examples will be available once the Rust implementation is complete.

Performance Analysis

Related Indicators