MeanReversionScore Node
Composite 0-100 mean reversion strength indicator
Overview
MeanReversionScore aggregates multiple statistical tests into a single 0-100 score indicating how strong mean reversion is. Score = 100 means very strong reversion; Score = 0 means trending behavior. It combines Hurst Exponent, Autocorrelation, Half-Life, and Variance Ratio into one interpretable metric.
Instead of juggling multiple indicators, traders get one clear signal: "Is this market reverting (90+) or trending (10-)?" This composite score is more robust than individual measures and adapts to market regime changes. High scores validate mean reversion entry signals. Low scores warn of trend changes.
Formula & Calculation
AC_score: Autocorr < -0.1 → Normalization to 0-100
HL_score: Half-Life < threshold → Higher score = faster reversion
VR_score: Variance Ratio < 1 → Higher score = more reverting
Score 40-70: Moderate/uncertain
Score < 40: Trending market (avoid MR)
Parameters
| Parameter | Default | Description |
|---|---|---|
| lookback | 50 | Period for component calculations |
| weights | Equal | Weight Hurst/AC/HL/VR equally or custom |
Common Use Cases
1. Entry Validation
Enter mean reversion only when MRS > 70. Skip entries when MRS < 40. Filters out unprofitable attempts in trending markets.
2. Strategy Switching
High MRS: Use mean reversion. Low MRS: Switch to momentum. Adapts entire strategyology to regime automatically.
3. Position Sizing
Size = Base × (MRS / 100). High reversion score = larger positions. Low score = reduce or sit out.
4. Exit Signals
MRS declining = exit early. Score drops below 50 = close positions. Predicts regime shift before P&L inflection.
Advantages & Limitations
Advantages
- Single interpretable 0-100 score
- Robust (multiple component agreement)
- Adapts to regime changes
- Directly validates MR setups
Limitations
- Component conflicts (rare but possible)
- Doesn't handle structural breaks
- Lookback dependency (lag)
- No directional information