Cybernetic Oscillator

The Cybernetic Oscillator is a zero-centered, energy-normalized momentum oscillator inspired by John Ehlers’ digital signal processing work. It combines a high-pass filter to remove drift, a 2nd-order Super Smoother to reduce noise, and RMS normalization so the amplitude remains comparable across markets and timeframes. The result is a responsive, clean oscillator that highlights direction changes and strength without being dominated by raw volatility.

Key Features at a Glance

• Band-pass behavior: High-pass → Super Smoother low-pass chain isolates tradable swings.
• RMS normalization: Divides by rolling RMS of the filtered signal to stabilize amplitude.
• Zero-centered: Clear polarity around the 0-line for bias flips.
• Threshold bands (±1 by default): Optional gating for “strong” moves.
• Polarity coloring: Green above zero, red below, for instant read.

Intuition First: The Signal Pipeline

Price → remove trend with a 2nd-order High-Pass → reduce noise via 2nd-order Super Smoother → scale by RMS of recent energy → plot oscillator with zero line and ±thresholds.

• The HP filter strips slow drift so momentum flips are timely.
• The Super Smoother suppresses jitter without excessive lag.
• RMS normalization keeps values consistent, making ±1 a meaningful, portable reference.

Formula Breakdown

1. High-Pass (2nd order)
   • Frequency set by hpPeriod. Coefficients use exp, cos and π to emulate a digital HP.
   • Output: hp, removing low-frequency drift.
2. Super Smoother (2nd order Low-Pass)
   • Frequency set by lpPeriod. Applied to the average of hp and hp[1] to reduce phase noise.
   • Output: lp, a smooth band-passed signal.
3. RMS Normalization
   • Compute rolling RMS over rmsLen of lp².
   • Normalize: osc = lp / max(rms, ε) so amplitude ≈ 1 in steady conditions.
4. Display & Thresholds
   • Plot osc, zero line, and optional ±threshold.
   • Dynamic color: green above zero, red below.

Inputs & Defaults

• hpPeriod = 30
  Controls how aggressively drift is removed.
  • Shorter (15–25): faster bias flips, more whipsaws.
  • Longer (35–50): steadier bias, slower to turn.
• lpPeriod = 20
  Governs smoothing of the high-pass output.
  • Shorter (10–16): snappier but noisier.
  • Longer (22–30): calmer but laggier.
• rmsLen = 100
  Window for energy normalization.
  • Shorter (50–80): amplitude adapts quickly to volatility shifts.
  • Longer (120–200): more stable scaling across regimes.
• threshold = 1
  Optional gating for “strong” moves.
  • Raise to 1.25–1.5 in choppy markets to reduce false positives.
  • Lower to 0.7–0.9 to be more permissive.

How to Read the Oscillator

• Zero-line flips: Crossing above 0 suggests bullish momentum bias; below 0 suggests bearish.
• Threshold zones: Moves beyond +threshold or below -threshold indicate stronger impulses.
• Divergences: Price making new extremes while osc fails to confirm often precedes momentum fades.
• Regime awareness: In strong trends, pullbacks may stall near the zero line; in ranges, price mean-reverts around zero more frequently.

Tip: Treat ±threshold as context, not absolute signals. Combine with trend/volatility filters.

Trading Playbooks (Examples, Not Advice)

1. Momentum Flip
   • Entry: Bias flip through zero in trend direction (e.g., price above a slow MA, oscillator crosses > 0).
   • Exit: First loss of momentum back through zero or trail with ATR.
   • Use threshold: If the first push also clears +threshold, allow a wider stop; otherwise, be tighter.
2. Pullback Continuation
   • Uptrend: After a dip above -threshold, look for osc to curl back up and cross zero.
   • Downtrend: Mirror logic with +threshold.
   • Idea: Avoid deeper counter-trend swings that pierce the opposite threshold.
3. Range Mean-Reversion
   • In sideways regimes, fade moves that tag ±threshold when price is at range extremes.
   • Exit near the zero line or mid-range.

Risk management: Always define invalidation (e.g., ATR-based stops or swing points) and position sizing before entries.

Best Practices & Common Pitfalls

• Don’t over-filter: Very long hpPeriod + lpPeriod can mute useful signals.
• Keep RMS realistic: Too small rmsLen can inflate amplitude after brief shocks.
• Mind regime shifts: Re-tune threshold when volatility regime changes.
• Use confirmations: Pair zero-cross ideas with a simple regime filter (e.g., MA slope) and a volatility filter (e.g., ATR percentile) to decide whether to run momentum or mean-reversion playbooks.

Code (ProBuilder)

Share this