How MIZAN weighs a setup.
Review the market conditions, the setup evidence, and the proposed reversal risk together. MIZAN™ makes those checks visible on your chart, alongside your own trading rules.
Three model families support that read, with order-flow evidence added in Institutional. The approach is described below; the calibration remains private.
Different markets. The same questions.
MIZAN is designed for stocks, indices, futures, forex and crypto, on intraday and longer timeframes. Read the conditions, inspect the evidence, and review the risk alongside your own method.
The instrument and timeframe still matter. Check the feed, loaded history, session filters and settings. Volume may be unavailable or represent price updates rather than traded quantity. Order-flow features depend on the available data; one configuration is not a universal fit.
TradingView’s volume-data notes ↗ · How MIZAN uses order-flow data →
01 · Adaptive equilibrium modeling
A self-correcting fair-price engine that fuses multiple volatility models in real time, weighted by which has been tracking price more closely rather than averaged.
Band width adapts to estimated volatility. With volatility anchoring enabled, selected volatility inputs use earlier bars. The center, model weighting and other width adjustments can still respond to current conditions.
Model influence is a normalized share of recent tracking accuracy, exponentially smoothed, and the fused center weights each model by the inverse of its smoothed error variance — influence is earned across bars, never seized by one.
02 · Multi-regime validation
Every setup is tested against the current market state. A fade that looks plausible in a range can fail the engine’s checks in a trend. Same shape. Different context.
The engine reads which game is being played before it scores a move, and a setup that fails a gate is refused and marked with the gate that stopped it — a named output on the chart, not a silent absence.
03 · Evidence-based scoring
No directional price forecasts, and no implied probability of profit. The engine estimates state and volatility only to normalize the conditions a setup is judged under; it does not predict where price will go.
The 0–100 score summarizes reversal evidence under the current settings. It is not a win probability. With bar-close confirmation enabled, signals confirm at the close; live readings remain provisional.
The score combines weighted observations of the setup and its market conditions.
Those observations cover band position, momentum extension, participation, model agreement, regime, rejection quality, wick structure, divergence, higher-timeframe alignment, volatility expansion, cross-market confirmation, liquidity sweeps, and level proximity.
Some individual contributions are capped, and the final score is limited to 0–100. Related observations can overlap; the number is not a count of independent confirmations.
Active presets apply score and reward-to-risk minimums. Custom settings can change those checks, while positive reversal geometry remains required. A high score does not override another failed requirement.
The score explains the setup. The gate decides admissibility. The block is the feature.
04 · Order-flow confluence INSTITUTIONAL TIER
Three data tiers keep the source of each read explicit: native footprint, intrabar reconstruction, and structural delta.
On supported TradingView plans, the engine reads the native footprint feed. When that data is unavailable, it reconstructs each bar’s footprint from lower-timeframe intrabars. The structural delta model sits beneath both.
When the native feed is live, the reconstruction is rescaled to its totals so the tiers report on the same scale.
The panel identifies the available data source. Reversal webhook payloads include source information; payload fields vary by event. Missing data does not become invented evidence.
TradingView classifies native footprint volume from intrabar price movement. The displayed ladder rows in the reviewed Institutional build remain reconstructed estimates, not guaranteed raw exchange aggressor data.
With bar-close confirmation enabled, the decision is confirmed at the close. Historical recalculation can still reflect changes in available data, loaded history, or settings; the active data tier stays visible.
TradingView’s footprint data documentation ↗
Institutional can compare logged reversal groups with and without order-flow evidence. These are descriptive groups, not a controlled experiment.
Additional evidence is not a promised outcome upgrade. Core and Institutional also have different defaults and calibration, so their signals need not match.
What the checks are looking for
The homepage groups the decision into market, setup and risk. Here are the technical families behind those checks. They are not seven equally weighted or statistically independent votes.
01Mean-Reversion Validation
regime-conditioned mean-reversion test · adaptive lookback
Confirms whether price historically responds to extension here — or whether momentum continuation is dominating.
Plain EnglishA stretched market is a rubber band. This check asks one question: at this distance, does price usually snap back here, or keep going?
02Trend Regime Detection
state classification: trend / chop / expansion
Separates directional expansion from sideways chop, so a fade can be blocked when the detected conditions do not support it.
Plain EnglishA trending market marches in one direction. A choppy market bounces inside a box. The same setup means opposite things in each, so MIZAN reads which one you are in first.
03Smart-Money Flow Analysis
one-sided pressure & aggressive-participation detection
Detects aggressive one-sided pressure and helps filter fades against it.
Plain EnglishIf selling keeps meeting each bounce, the pressure matters. This check weighs it before a countertrend setup qualifies. It does not identify who is trading.
04Adaptive Equilibrium Engine
multi-model volatility fusion · live weighting
A self-correcting balance model that continuously recalibrates the system’s center of gravity.
Plain EnglishThis is the fair-price line the bands hang from. It moves with the market instead of sitting on a fixed setting.
05Regime-Shift Detection
large-move detection · stricter score requirement
After an unusually large price move, the engine can require a higher score before a reversal passes.
Plain EnglishA sudden move can raise the bar for the next reversal. Scheduled news windows are configured separately; MIZAN does not automatically read a news calendar.
06Market Roughness Filter
price-path roughness estimation · noise suppression
Estimates price-path roughness and structural clarity, which influence the adaptive bands and supporting market read.
Plain EnglishSome price action moves like stairs. Some moves like a jackhammer. That distinction informs the read; noise alone does not determine the final score.
07Dynamic Risk Calibration
per-signal stop & target geometry · R:R gate · failure lockouts
Calculates entry, stop, target and reward-to-risk for reversal candidates. It marks candidates BLOCKED when the configured risk requirements fail, and can temporarily block another signal after a failure at the same level.
Plain EnglishQualified reversal signals have entry, stop and target references. Continuation marks are separate and provide a target only. No room to be paid? Blocked. Just failed here? Locked out until the tape resets.
Where price action fits
A familiar candle or a stretched price is a candidate, not a complete decision. MIZAN adds context to the patterns you already read; it does not replace your trading rules.
Bollinger Bands describe dispersion and RSI describes momentum. MIZAN combines several observations with explicit setup requirements. Read the mechanics comparison.
A description is not a forecast
The models estimate current state and volatility. Neither a status nor a score is a promise about the next bar. Passing the checks, later price movement, and a trader’s actual outcome are different things.
What we do not disclose
The exact parameter sets. The fusion weighting logic. The calibration thresholds. The proprietary extensions to standard models.
This is intentional, and it is the whole boundary. Everything above describes what the engine decides and why, which is enough to judge the approach before deciding whether the numbers are worth paying for. The vocabulary is public. The calibration is ours.
What it cannot do
It does not forecast price direction. It estimates market state and volatility, then grades the evidence already on the chart.
The current Institutional script requires TradingView Premium or Ultimate. Lower-tier accounts cannot run it. Fallback handles missing footprint data on an eligible account; it does not bypass that requirement.
Reconstruction quality falls on thin instruments where intrabar volume is sparse.
The ledger models reversal outcomes from loaded chart history; continuations are excluded. Costs default to zero. Experimental hourly-outcome feedback is a separate opt-in and is off by default.
By default, a bar that touches both stop and target is recorded as a loss. Institutional offers optional intrabar resolution that can establish target-first only with adequate, unambiguous coverage. That option is off by default.
On authorship
The evidence families and footprint concepts named above are public methodology. The implementation is not borrowed: every line is written from scratch, and no code in it is taken from another author’s script, on any platform.