Entropy Momentum Filter [EMF] -FibonacciFlux

Entropy Momentum Filter [EMF]

What it does
Measures directional conviction in price action using Shannon entropy of log returns. Low entropy means strong directional conviction — the market is trending. High entropy means indecision — the market is ranging or chaotic.

How it works

• Entropy Calculation: Builds a histogram of log returns over a rolling lookback window, then computes Shannon entropy H = -sum(p * ln(p))
  
• Normalization: Divides by ln(N) to produce a 0-1 scale where 0 = perfect order and 1 = maximum randomness
  
• Color Gradient: Green (low entropy, conviction) → Gray (neutral) → Red (high entropy, indecision)
  
• Reference Lines: 0.3 low entropy threshold, 0.5 midpoint, 0.7 high entropy threshold
  

Key Features

• Single clean concept — Shannon entropy applied to price changes
  
• Normalized 0-1 output — no parameter-dependent interpretation needed
  
• Built-in alert for conviction spikes (entropy drops below 0.3)
  
• Only 2 inputs: lookback period and histogram bin count
  
• Lightweight — minimal computational overhead
  

How to use

• Conviction filter: Only take trend-following trades when entropy is below 0.5
  
• Range detection: High entropy (>0.7) suggests avoid trend strategies or switch to mean-reversion
  
• Conviction spike alert: Built-in alert fires when entropy drops below 0.3 — strong move likely starting
  
• Combine with directional indicators: Use EMF as a filter on top of your existing signals
  

Inputs

• Lookback period (default 20) — rolling window for entropy calculation
  
• Histogram bins (default 10) — granularity of the return distribution
  

Works on all timeframes and instruments. Particularly effective as a regime filter combined with trend-following or mean-reversion strategies.