Indicators Description

1. Component Deconstruction¶

This section dissects each individual indicator and calculation engine within the script, detailing its configuration and any mathematical modifications.

Volatility & Sensitivity Engine¶

• Average True Range (ATR)

  • Specific Configuration: ta.atr(atrLength) with a default atrLength of 14.

  • Functional Modification: The raw ATR value (atrVal) is used as the base unit for risk management (Stop Loss/Take Profit). It is also fed into an adaptive sensitivity modifier.

• Adaptive Sensitivity Multiplier (adaptiveMult)

  • Specific Configuration: Uses ta.percentrank(atrVal, 100) to rank the current ATR value against its last 100 values.

  • Functional Modification: This is a custom volatility regime filter. It modifies the final Confidence Score based on market volatility:

    • High Volatility (atrPctRank > 70): adaptiveMult becomes 0.7. This reduces the confidence score, making signals harder to trigger. The logic is that high volatility produces more noise and false signals, requiring a higher conviction threshold.

    • Low Volatility (atrPctRank < 30): adaptiveMult becomes 1.3. This inflates the confidence score, making signals easier to trigger. The logic is to increase sensitivity in quiet markets where significant moves are less frequent.

    • Normal Volatility: adaptiveMult is 1.0 (no change).

    • The user can override this with fixed “Low” (0.7) or “High” (1.3) settings.

Volume & Momentum Analysis¶

• Synthetic Volume Delta (CVD)

  • Specific Configuration: This is a custom, non-standard calculation. It does not use request.security() to pull tick data but rather approximates delta from bar data.

  • Functional Modification:

    1. Close Position Ratio: closePosition = (close - low) / (high - low) calculates the closing price’s position within the bar’s range (0 for closing at the low, 1 for closing at the high).

    2. Volume Apportionment: buyVolume = volume * closePosition and sellVolume = volume * (1.0 - closePosition). This model assumes buying pressure is proportional to how high the bar closes, and vice-versa.

    3. Bar Delta: volumeDelta = buyVolume - sellVolume.

    4. Session Accumulation: sessionCVD accumulates volumeDelta throughout a single trading day (isNewSession = ta.change(time("D")) != 0), resetting at the start of each new day. This provides a session-specific view of net buying/selling pressure.

    5. Smoothing: The raw sessionCVD is smoothed using ta.ema with a default length of 5 (cvdSmoothing) to reduce noise and identify the underlying trend in order flow.

• CVD Divergence

  • Specific Configuration: Uses ta.pivotlow and ta.pivothigh with a lookback of 10 (swingLookback) to identify price swing points.

  • Functional Modification: It implements a classic divergence detection engine:

    • Bullish Divergence: Triggers true if a new price pivot low is lower than the previous one, BUT the corresponding cvdSmoothed value at the new pivot is higher than at the previous one.

    • Bearish Divergence: Triggers true if a new price pivot high is higher than the previous one, BUT the corresponding cvdSmoothed value is lower. This is a non-standard implementation as it uses ta.pivotlow(-high) to find high pivots.

• Pressure Index

  • Specific Configuration: ta.ema(volumeDelta / math.max(volume, 1) * 100, cvdSmoothing).

  • Functional Modification: This oscillator normalizes the volumeDelta by the total volume of the bar, creating a percentage-based measure of net pressure. It is then smoothed with an EMA (default length 5). It is designed to be a more sensitive, lag-reduced indicator of immediate buying or selling pressure compared to the cumulative CVD.

Event-Based Pattern Detection¶

• Liquidity Sweep

  • Specific Configuration: Uses ta.pivothigh and ta.pivotlow with a lookback of 10 (swingLookback) to establish recent swing points.

  • Functional Modification: A sweep is a multi-condition boolean event. A bearishSweep requires:

    1. Price pierces the most recent swing high (high > recentSwingHigh).

    2. Price fails to hold the level, closing back below it (close < recentSwingHigh).

    3. The candle is bearish (close < open).

    4. The event occurs on high volume (volume > ta.sma(volume, 20) * sweepThreshold). The logic for bullishSweep is the inverse.

• Order Absorption

  • Specific Configuration: Compares current volume to a 20-period SMA of volume.

  • Functional Modification: Identifies candles indicative of large passive orders absorbing aggressive market orders. It requires:

    1. High Volume: volume > avgVolume * absorbVolMult (default: > 2.0x average).

    2. Indecision: The candle’s body is a small fraction of its total range ((bodySize / barRange) < absorbBodyRatio, default: < 30%). This signifies a battle between buyers and sellers with no clear winner, despite high volume.

Market Structure & Regime Filters¶

• Institutional VWAP (Volume-Weighted Average Price)

  • Specific Configuration: This is a session-based VWAP, manually calculated to reset daily. It also calculates the session’s standard deviation of price from the VWAP.

  • Functional Modification:

    • VWAP: sumPV / sumVol.

    • Standard Deviation: math.sqrt(sumPV2 / sumVol - vwapVal * vwapVal). This is the correct mathematical formula for population standard deviation for a weighted dataset.

    • Bands: The standard deviation is multiplied by user-defined values (vwapBandMult1, vwapBandMult2) to create deviation bands, which act as dynamic support/resistance zones.

• Point of Control (POC)

  • Specific Configuration: ta.sma(typicalPrice * volume, pocLength) / ta.sma(volume, pocLength). Default pocLength is 20.

  • Functional Modification: This is not a traditional volume profile POC. It is a moving VWAP with a fixed lookback period of 20 bars, making it a shorter-term measure of the volume-weighted price center.

• Choppiness Index

  • Specific Configuration: Standard implementation with a default chopLength of 14.

  • Functional Modification: Calculates 100 * log10(ATR_Sum / HL_Range) / log10(Length). A value above the chopThreshold (default 65.0) defines the market as “choppy,” deactivating all trade signals. This acts as a master “regime filter.”

• Session Awareness

  • Specific Configuration: Uses the time() function to check if the current bar falls within user-defined session windows (NY, London).

  • Functional Modification: Creates a boolean flag inKillZone. This acts as a gatekeeper, only allowing signals during these specified high-liquidity periods. It also applies a sessionMult of 1.2 to the Confidence Score during these times, increasing the likelihood of a signal.

2. Logic Layering & Confluence¶

The script’s engine avoids reliance on any single indicator by layering multiple conditions in a hierarchical structure.

• Interaction Dynamics: The core logic is built on Confluence and Threshold Crosses. It does not simply look for indicators to agree, but quantifies the degree of agreement through a scoring system.

• Hierarchical Filtering: The signal generation follows a strict, multi-stage filtering process:

  1. Stage 1: The “Pillars” (Confluence Check):

     • The script defines 7 distinct bullish and 7 bearish “pillars” of evidence (e.g., CVD Divergence, Liquidity Sweep, VWAP Bounce).

     • A trade setup is only considered if a minimum of 2 pillars are present (bullPillarCount >= 2). This is the first and most crucial confluence requirement. A single event, like a sweep, is insufficient on its own.

  2. Stage 2: The “Bias” (Directional Check):

     • A separate, weighted scoring system (bullishBias, bearishBias) determines the dominant market pressure. Conditions deemed more significant (e.g., a Sweep or Divergence) are given 2 points, while others (e.g., EMA alignment) receive 1.

     • The trigger requires the bias to align with the trade direction (e.g., bullishBias > bearishBias for a long). This prevents taking long trades in an overwhelmingly bearish environment, even if two bullish pillars are present.

  3. Stage 3: The “Confidence Score” (Strength Check):

     • This is a granular scoring system (0-100) where different components contribute points to a rawConfidence total (e.g., a CVD divergence adds 20 points, an absorption event adds 6).

     • This score is then dynamically adjusted by the adaptiveMult (volatility) and sessionMult (time of day).

     • The final confidence must exceed the minConfidence threshold (default 30). This acts as a quantitative filter for signal quality.

  4. Stage 4: The “Gatekeepers” (Regime & State Check):

     • Even if all the above conditions are met, a signal is only generated if it passes three final boolean checks:

       • passesChop: The market must not be in a choppy state.

       • passesSession: The trade must occur within a designated “Kill Zone.”

       • posState == 0: The system must be flat (no active trade). This prevents pyramiding.

       • signalCooldown: A minimum of 3 bars must pass between signals.

3. The Execution Engine¶

Trigger Logic¶

The final boolean logic for a long entry (longEntry) is the culmination of the entire filtering hierarchy:

longSignal = (bullishBias > bearishBias) and (bullPillarCount >= 2) and (confidence >= minConfidence) and (not isChoppy) and (inKillZone) and barstate.isconfirmed

longEntry = longSignal and (bar_index - lastSignalBar >= 3) and (posState == 0)

• bullishBias > bearishBias: The weighted directional evidence must be bullish.

• bullPillarCount >= 2: At least two distinct bullish patterns must have occurred.

• confidence >= minConfidence: The quantitatively measured signal strength must pass the minimum threshold.

• not isChoppy and inKillZone: The market regime and session time must be permissive.

• signalCooldown and posState == 0: State management rules must be satisfied.

Exit Logic¶

The script employs a hybrid exit strategy with both static and dynamic conditions.

• Static Exits (Risk Management):

  • Stop Loss: close - atrVal * atrMultSL (for longs). The stop is placed at a distance determined by a multiple of the ATR at the time of entry.

  • Take Profit: close + atrVal * atrMultTP (for longs). The target is also based on the entry-time ATR.

  • Mathematical Constants:

    • atrMultSL (default 1.5): A multiplier of 1.5 means the initial risk is 1.5 times the 14-period ATR. A smaller value implies a tighter stop and higher risk of being stopped out by noise.

    • atrMultTP (default 1.0): A multiplier of 1.0 sets the initial reward target at 1x the 14-period ATR. With default settings, this establishes an initial Risk-to-Reward ratio of 1.5 : 1.0, meaning the potential loss is greater than the potential reward. This profile is typical of high-win-rate scalping strategies where the goal is to capture small, probable moves.

• Dynamic Exit (Momentum-Based):

  • longMomExit = posState == 1 and momentumBearish and pressureIndex < -5

  • This condition allows for an early exit if momentum shifts against the position before the SL or TP is hit. It requires:

    1. A bearish divergence between price ROC and CVD ROC (momentumBearish).

    2. The immediate selling pressure (pressureIndex) to cross a hard-coded threshold of -5.

  • This acts as a mechanism to protect profits or cut losses early if the reversal thesis of the trade entry fails to materialize.