OPEN-SOURCE SCRIPT
已更新 Auction Regime Router Entropy Gate & Hurst Memory
Auction Regime Router — Entropy Gate & Hurst Memory
What it is
Every structure playbook fails in the wrong regime. Fading the value-area edge works when price is anti-persistent (stretches snap back); riding a breakout works when price is persistent (moves feed on themselves); and nothing structural works when the tape is noise. This tool measures two things — how much structure exists, and what kind it is — and routes to a plain-language answer: FADES VIABLE / BREAKOUTS VIABLE / STAND ASIDE. It decides which of your tools to trust, never buy or sell.
The two measurements (and how they work together)
Permutation entropy (Bandt–Pompe 2002) — the gate. It measures how disordered the recent price sequence is from the frequencies of ordinal patterns (which of the 6 orderings each price triplet takes). High entropy = all patterns equally likely = noise = no structural edge. The gate is self-calibrated: entropy is ranked against its own recent history, so "noisy" means noisy for this symbol and timeframe.
Hurst exponent (Hurst 1951; Mandelbrot) — the router. Memory via diffusion scaling: how the dispersion of K-bar returns grows with K. H > 0.5 = persistent → continuation regime; H < 0.5 = anti-persistent → reversion regime. Research supports the routing: mean reversion is empirically more probable and faster during anti-persistent periods.
The mashup logic is a hierarchy, not a mixture: the entropy gate overrides the Hurst read. If the tape is noise, the router says STAND ASIDE regardless of what H says — because a memory estimate on noise is meaningless.
The honesty steps
A dead zone around H = 0.5 (default 0.45–0.55): near a random walk the memory read is unreliable, so the router says MIXED rather than pretending. Practitioners commonly require a margin before activating a playbook; both thresholds are inputs.
A minimum-dwell filter (the standard anti-chattering design from switched-systems control): a new regime is announced only after it survives a set number of confirmed bars, so the read doesn't flip-flop bar to bar. The cost is that many bars of lag — stated, and adjustable.
Estimates are proxies from bar data with overlapping windows — descriptive of the recent past, not a prediction. The dashboard shows the state, how long it has persisted (regime age), how dominant it has been recently (stability %), and any pending regime with a countdown — nothing more.
How to use it
Add to any liquid symbol/timeframe; defaults suit intraday index futures. The script requests no external data of any kind, so it runs on every plan and every symbol.
Glance at the regime lane — the thin colored strip at the bottom of the pane: blue = continuation, violet = reversion, amber = noise, gray = mixed. The palette is deliberately direction-neutral — no green or red anywhere in regime coding, so nothing can be misread as a buy or sell.
The HTF STACK row shows the raw regime on three higher timeframes derived as multiples of the chart (defaults 3×, 5×, 15× — so a 5m chart reads 15m/25m/75m automatically, adapting to any chart). A ✓ in green = every timeframe agrees on the same actionable regime (strongest context). A ⚠ in amber = a higher timeframe reads NOISE or the opposite regime while the chart claims a playbook (weakest — reduce or wait).
Read the dashboard for detail: REVERSION → your value-area fade / band-reversion tools are in their element; CONTINUATION → your breakout / drive tools are; NOISE → the gate is closed, stand aside; MIXED → no clear routing, reduce. STABILITY shows how settled the read is; PENDING shows a forming regime with a countdown.
Regime-change tags print only on announced (dwell-confirmed) changes; alerts fire on entering each state.
Best used as the selector above your structure toolkit rather than as a standalone display.
What makes it original
Hurst and entropy oscillators exist. What this adds: (1) the hierarchy — a self-calibrated entropy gate that can veto the memory read, instead of two numbers side by side; (2) routing to auction playbooks in plain language (fade vs breakout viability), not a raw statistic; (3) honest dead zones, a minimum-dwell announcement filter, and stability/pending context instead of a binary flip at H = 0.500. It is a decision-hygiene tool for structure traders.
Concept credits
Ordinal-pattern (permutation) entropy — C. Bandt & B. Pompe (2002). Long-memory / rescaled-range analysis — H. E. Hurst (1951); fractal market framing — B. Mandelbrot. Regime-gated strategy selection — standard quantitative practice. Implementation and charting design are the author's own.
Important disclaimer
Research and education only. Not financial advice, not a signal service, not a guarantee of future results. Regime labels are descriptive statistics of recent bars; regimes change without warning and estimates are proxies. Validate independently and manage your own risk.
What it is
Every structure playbook fails in the wrong regime. Fading the value-area edge works when price is anti-persistent (stretches snap back); riding a breakout works when price is persistent (moves feed on themselves); and nothing structural works when the tape is noise. This tool measures two things — how much structure exists, and what kind it is — and routes to a plain-language answer: FADES VIABLE / BREAKOUTS VIABLE / STAND ASIDE. It decides which of your tools to trust, never buy or sell.
The two measurements (and how they work together)
Permutation entropy (Bandt–Pompe 2002) — the gate. It measures how disordered the recent price sequence is from the frequencies of ordinal patterns (which of the 6 orderings each price triplet takes). High entropy = all patterns equally likely = noise = no structural edge. The gate is self-calibrated: entropy is ranked against its own recent history, so "noisy" means noisy for this symbol and timeframe.
Hurst exponent (Hurst 1951; Mandelbrot) — the router. Memory via diffusion scaling: how the dispersion of K-bar returns grows with K. H > 0.5 = persistent → continuation regime; H < 0.5 = anti-persistent → reversion regime. Research supports the routing: mean reversion is empirically more probable and faster during anti-persistent periods.
The mashup logic is a hierarchy, not a mixture: the entropy gate overrides the Hurst read. If the tape is noise, the router says STAND ASIDE regardless of what H says — because a memory estimate on noise is meaningless.
The honesty steps
A dead zone around H = 0.5 (default 0.45–0.55): near a random walk the memory read is unreliable, so the router says MIXED rather than pretending. Practitioners commonly require a margin before activating a playbook; both thresholds are inputs.
A minimum-dwell filter (the standard anti-chattering design from switched-systems control): a new regime is announced only after it survives a set number of confirmed bars, so the read doesn't flip-flop bar to bar. The cost is that many bars of lag — stated, and adjustable.
Estimates are proxies from bar data with overlapping windows — descriptive of the recent past, not a prediction. The dashboard shows the state, how long it has persisted (regime age), how dominant it has been recently (stability %), and any pending regime with a countdown — nothing more.
How to use it
Add to any liquid symbol/timeframe; defaults suit intraday index futures. The script requests no external data of any kind, so it runs on every plan and every symbol.
Glance at the regime lane — the thin colored strip at the bottom of the pane: blue = continuation, violet = reversion, amber = noise, gray = mixed. The palette is deliberately direction-neutral — no green or red anywhere in regime coding, so nothing can be misread as a buy or sell.
The HTF STACK row shows the raw regime on three higher timeframes derived as multiples of the chart (defaults 3×, 5×, 15× — so a 5m chart reads 15m/25m/75m automatically, adapting to any chart). A ✓ in green = every timeframe agrees on the same actionable regime (strongest context). A ⚠ in amber = a higher timeframe reads NOISE or the opposite regime while the chart claims a playbook (weakest — reduce or wait).
Read the dashboard for detail: REVERSION → your value-area fade / band-reversion tools are in their element; CONTINUATION → your breakout / drive tools are; NOISE → the gate is closed, stand aside; MIXED → no clear routing, reduce. STABILITY shows how settled the read is; PENDING shows a forming regime with a countdown.
Regime-change tags print only on announced (dwell-confirmed) changes; alerts fire on entering each state.
Best used as the selector above your structure toolkit rather than as a standalone display.
What makes it original
Hurst and entropy oscillators exist. What this adds: (1) the hierarchy — a self-calibrated entropy gate that can veto the memory read, instead of two numbers side by side; (2) routing to auction playbooks in plain language (fade vs breakout viability), not a raw statistic; (3) honest dead zones, a minimum-dwell announcement filter, and stability/pending context instead of a binary flip at H = 0.500. It is a decision-hygiene tool for structure traders.
Concept credits
Ordinal-pattern (permutation) entropy — C. Bandt & B. Pompe (2002). Long-memory / rescaled-range analysis — H. E. Hurst (1951); fractal market framing — B. Mandelbrot. Regime-gated strategy selection — standard quantitative practice. Implementation and charting design are the author's own.
Important disclaimer
Research and education only. Not financial advice, not a signal service, not a guarantee of future results. Regime labels are descriptive statistics of recent bars; regimes change without warning and estimates are proxies. Validate independently and manage your own risk.
發行說明
Minor change發行說明
v1.1 — Finishing pass (no engine change)- Cleared a Pine warning (CW10004): the smoothed Hurst read called ta.ema from inside a "skip while
warming up" check (chart and higher-timeframe). It now runs every bar via a seed-on-first-valid
EMA that reproduces the exact same output after warm-up — only the warning is gone.
- Added a Data Window (EXP_) export bus so the entropy rank, Hurst exponent and regime state can
feed other scripts via input.source().
- Added the MPL-2.0 licence header.
- No change to the entropy gate, the Hurst memory, the regime routing, the multi-timeframe reads or
the dashboard — everyday behaviour is identical.
Descriptive regime-gating tool, not investment advice.
開源腳本
秉持TradingView一貫精神，這個腳本的創作者將其設為開源，以便交易者檢視並驗證其功能。向作者致敬！您可以免費使用此腳本，但請注意，重新發佈代碼需遵守我們的社群規範。
免責聲明
這些資訊和出版物並非旨在提供，也不構成TradingView提供或認可的任何形式的財務、投資、交易或其他類型的建議或推薦。請閱讀使用條款以了解更多資訊。
開源腳本
秉持TradingView一貫精神，這個腳本的創作者將其設為開源，以便交易者檢視並驗證其功能。向作者致敬！您可以免費使用此腳本，但請注意，重新發佈代碼需遵守我們的社群規範。
免責聲明
這些資訊和出版物並非旨在提供，也不構成TradingView提供或認可的任何形式的財務、投資、交易或其他類型的建議或推薦。請閱讀使用條款以了解更多資訊。