Evaluating 10 fractal / multifractal / chaos / complexity candidate features for whether each adds information orthogonal to the live production panel — one candidate per overnight firing, read-only across 10 regime slices × 21 (symbol, threshold) cells on ClickHouse. Gate: ‖ρ‖≤0.85 ∧ R²≤0.85 ∧ h_norm≥0.05 (worst-valid-cell). Scope: orthogonality only — no null, no signal/target. Resource-capped to 2 cores @ 200% / 4 GB, spare capacity only.
| # | Candidate | Verdict | Max ‖ρ‖ | Valid cells | Cost |
|---|---|---|---|---|---|
| 01 | 🕸️ Ordinal partition network — global node entropy network-complexity · ordpy | WATCH | 0.94 | 207 | medium |
| 02 | 🪢 Higuchi fractal dimension fractal-dimension · antropy | PASS | 0.85 | 207 | fast |
| 03 | 📏 Generalised Hurst H(q=2) hurst-multifractal · nolds | WATCH | 0.87 | 207 | medium |
| 04 | 🎼 Permutation Jensen–Shannon distance to 1/f² noise complexity-entropy · ordpy | WATCH | 0.90 | 207 | fast |
| 05 | 🌀 MF-DFA multifractal spectrum width Δα (price) multifractal · custom | PASS | 0.56 | 207 | slow |
| 06 | 📊 Δh(q) on |returns| — volatility multifractality multifractal-volatility · MFDFA | PASS | 0.68 | 207 | slow |
| 07 | 📈 Δh(q) on returns multifractal · MFDFA | PASS | 0.65 | 207 | slow |
| 08 | ⚡ Δh(q) on volatility increments (Δσ²) multifractal-volatility · MFDFA | PASS | 0.71 | 207 | slow |
| 09 | 🌌 Correlation dimension D₂ (chaos) chaos-dimension · nolds | WATCH | 0.89 | 207 | slow |
| 10 | 🦋 Largest Lyapunov exponent (Rosenstein) chaos · nolds | PASS | 0.81 | 207 | slow |