MarketEntropy Node
Information content and disorder measurement
Overview
Market Entropy measures the degree of order or chaos in price movements using Shannon Entropy from information theory. High entropy = random/chaotic price action (low predictability). Low entropy = ordered/organized patterns (high predictability). It quantifies market efficiency: perfectly efficient = maximum entropy; mispriced = low entropy.
When entropy is low, patterns are exploitable. When entropy is high, noise dominates. This helps traders adapt strategy complexity: in high-entropy (random) markets, use simple strategies; in low-entropy (ordered) markets, use complex feature-rich strategies for edge.
Formula & Calculation
Sum over all bins of histogram
Parameters
| Parameter | Default | Description |
|---|---|---|
| lookback | 100 | Periods for entropy calculation |
| bins | 20 | Price histogram bins |
Common Use Cases
1. Market Predictability
Low entropy = more predictable patterns. Trade momentum/patterns. High entropy = random = tighten stops, reduce position size.
2. Regime Detection
Rising entropy = transition from ordered to chaotic. Falling entropy = transition from random to patterned (often precedes trends).
3. Model Complexity
Low entropy: Use complex ML models (they find patterns). High entropy: Use simple strategies (overfitting risk too high).
4. Position Sizing
Size inversely to entropy: High entropy (chaotic) = smaller positions. Low entropy (patterns clear) = larger positions.
Advantages & Limitations
Advantages
- Measures predictability directly
- Theoretically sound (info theory)
- Guides strategy selection
- Independent of price level
Limitations
- Bin selection affects results
- Slow to change (lookback lag)
- No directional information
- Requires adequate samples