Session Sweep Reversal Predictor - PhenLabs

The Session Sweep Reversal Predictor - PhenLabs indicator evaluates confirmed session liquidity sweeps to estimate whether price is more likely to reverse or continue. It tracks Asian, London, and New York session ranges and uses a rolling Bayesian model to calculate probabilities based on historical sweep outcomes and displacement quality.

Usage

The Usage section describes how the script can be used to identify high-probability reversal or continuation zones after a session level has been breached.

Session Liquidity Analysis

The tool automatically identifies the High and Low of the Asian, London, and New York sessions. When a later session breaches one of these levels, the script evaluates the nature of the "sweep." For example, the London session can sweep the Asian range, while the New York session can sweep either London or Asian levels.

Reversal vs. Continuation Probability

Each qualified sweep is analyzed based on candle structure, displacement (ATR expansion), and body-to-range quality.

• High Reversal Probability: Indicated by a green label when the probability exceeds the user-defined threshold. This suggests the sweep is likely a rejection, favoring a move back into the previous range.
• High Continuation Probability: Indicated by a red label when the probability exceeds the threshold. This suggests price intent to continue trending in the direction of the sweep.
• Gray Label: Appears when the Bayesian model does not find a significant statistical edge based on the recent lookback window.

Displacement Filtering

To reduce noise from minor price fluctuations, the indicator filters for displacement. Users can set a minimum candle size (multiplier of ATR) and a minimum body-to-range ratio to ensure only high-conviction sweeps are tracked and labeled.

Details

The script operates through a multi-stage process to ensure statistical relevance:

1. Range Tracking: Independently monitors and locks session extremes. Once a session ends, its high and low become the reference points for subsequent sessions.
2. Sweep Detection: Monitors for "Wick Sweeps" (rejections) or "Failed-Close Sweeps" (immediate reclaims after a close beyond the level).
3. Bayesian Modeling: A rolling Naive Bayes model stores the characteristics of the last X sweep events (defined by the Rolling Lookback). It categorizes these events into "reversals" or "continuations" based on price action over a fixed forward window.
4. Laplace Smoothing: Uses a smoothing factor (α) to prevent the model from becoming overconfident or producing extreme results when the sample size is small.
5. Outcome Classification: After the "Outcome Forward Bars" period, the script verifies if price moved at least one ATR back through the level (Reversal) or failed to do so (Continuation).

Settings

Sessions

• Session Timezone: Selects the timezone for hour-based session classification.
• Asian / London / New York Session: Defines the start and end times for each reference window.

Sweep Detection

• Detect Wick Sweeps: Enables detection of price straddling a level and closing back through it.
• Detect Failed-Close Sweeps: Enables detection where price closes beyond a level but is reclaimed by the following bar.
• Session Sweep Logic: Toggles allow/disallow specific session interactions (e.g., London sweeping Asian range).

Displacement Filters

• ATR Length: Sets the period for the Average True Range used in displacement calculations.
• Min Candle Size × ATR: The required magnitude of the sweep candle relative to volatility.
• Min Body/Range Ratio: Filters out weak wicks; requires the candle body to be a certain percentage of the total candle range.

Bayesian Model

• Rolling Lookback: The number of previous sweep events used to calculate current probabilities.
• Laplace Smoothing α: Adjusts the sensitivity of the probability calculations.
• Outcome Forward Bars: The number of bars the script looks ahead to classify a sweep as a reversal or continuation.
• Reversal/Continuation Threshold %: The probability required to trigger colored labels and alerts.

FAQ

How is a "reversal" defined in this model? A sweep is classified as a reversal if, within the "Outcome Forward Bars" window, price moves at least one ATR (measured at the time of the sweep) back across the swept level.

What do the different box colors represent? By default, blue boxes represent the Asian session range, and orange boxes represent the London session range. The background tint indicates which session is currently active.

How do I access the Session Sweep Reversal Predictor - PhenLabs? You can get access on the LuxAlgo Library for charting platforms like TradingView, MetaTrader (MT4/MT5), and NinjaTrader for free.