Indicators Description

1. Component Deconstruction¶

This analysis deconstructs the core mathematical and logical components of the TSI Adaptive Scalper v4.2 OVL-SP script.

A. True Strength Index (TSI)¶

The central oscillator of the strategy.

• Specific Configuration:

  • Price Source: close.

  • Long Period (fLong): Adaptive, determined by a matrix of Timeframe Class and Instrument Type. Ranges from 13 (M1/Metal) to 40 (D1/Forex).

  • Short Period (fShort): Adaptive, determined by the same matrix. Ranges from 7 (M1/Metal) to 20 (D1/Forex).

  • Signal Period (fSignal): Adaptive, determined by the matrix. An EMA of the final TSI value. Ranges from 5 to 18.

  • Overbought/Oversold Levels (lvlOB/lvlOS): Base levels are adaptive via the matrix (e.g., 25.0 for M1/Metal, 42.0 for D1/Index). These base levels are then dynamically adjusted.

• Functional Modification: The script employs two significant, non-standard modifications to the TSI calculation:

  1. Adaptive Post-Smoothing (tsiAdapt): The raw TSI value (tsiRaw) is subjected to a custom smoothing filter before being used as the final tsi value. The formula is tsiAdapt := activeAlpha * tsiRaw + (1.0 - activeAlpha) * nz(tsiAdapt[1], tsiRaw).

     • Logic: The smoothing factor, activeAlpha, is dynamically calculated based on the ratio of a fast ATR (10-period) to a slow ATR (40-period).

     • Intended Effect: During high volatility (high ATR ratio), activeAlpha increases, making the TSI more responsive and reducing lag. During low volatility, activeAlpha decreases, increasing the smoothing effect to filter out market noise. This creates a volatility-normalized momentum reading.

  2. Dynamic Volatility Thresholds (dynOB/dynOS): The OB/OS levels are not static. They are calculated by multiplying the base level (baseOB) by a volatility ratio (clampedVol).

     • Logic: clampedVol is the ratio of a 14-period ATR to a 50-period ATR, clamped between a floor (0.6) and a ceiling (1.6).

     • Intended Effect: In high-volatility regimes, the OB/OS bands widen, requiring a more significant price move to trigger an extreme reading. This reduces false signals during volatile expansion phases. In low-volatility regimes, the bands contract.

B. Exponential Moving Average (EMA)¶

Used as a primary trend filter.

• Specific Configuration:

  • Price Source: close.

  • Length (emaLenFinal): Adaptive, based on Timeframe Class. Ranges from 21 (M1) to 60 (D1).

• Functional Modification: None. It is a standard EMA used to establish a directional bias (close > emaVal for bullish).

C. Average Directional Index (ADX)¶

Used as a trend strength filter.

• Specific Configuration:

  • DI Length: 14 periods (adxLen).

  • ADX Smoothing Length: 14 periods (adxLen).

  • Threshold (adxThFinal): Adaptive, based on Timeframe Class. Ranges from 20.0 to 23.0.

• Functional Modification:

  • Double Smoothing: An optional EMA (adxSmoothLen = 5) is applied to the final ADX value. This is intended to reduce choppiness in the ADX line, providing a more stable reading of trend strength at the cost of minor additional lag.

D. Bollinger Bands (BB) / BB Width¶

Used as a consolidation and whipsaw filter.

• Specific Configuration:

  • Price Source: close.

  • Length (bbPeriod): 20.

  • Standard Deviations (bbStdDev): 2.0.

• Functional Modification: The script does not use the bands for signals. It calculates the Normalized Bollinger Band Width: (bbUpper - bbLower) / bbBasis * 100.0.

  • Logic: This value is compared to a threshold (consolThr) to determine if the market is in a state of consolidation. The threshold itself is adaptive, defaulting to the 25th percentile of the BB Width over the last 100 bars.

  • Intended Effect: This module acts as a regime filter. When isConsolidation is true, it penalizes the signal score and increases the signal cooldown period, systematically reducing exposure in low-probability, ranging markets.

E. Average True Range (ATR)¶

A multi-purpose volatility engine.

• Specific Configuration: Used in three separate contexts:

  1. TSI Reactivity: 10-period (fast) and 40-period (slow).

  2. Dynamic Thresholds: 14-period (fast) and 50-period (slow).

  3. TP/SL Calculation: 14-period (tpslATRPeriod).

• Functional Modification: ATR is never used as a direct signal. It serves exclusively as a quantitative measure of volatility to dynamically modify the behavior of other components (TSI, OB/OS levels, and risk parameters).

F. Relative Volume¶

A volume-based confirmation filter.

• Specific Configuration:

  • Source: volume.

  • Lookback Period: 20-period SMA (volRelPeriod).

  • Thresholds: Strong volume is > 1.3x the SMA; weak volume is < 0.6x the SMA.

• Functional Modification: This is not a binary filter but a scoring modifier. It adds a progressive bonus (volScoreAdj) for signals occurring on high volume and a progressive penalty for signals on low volume, directly influencing the final confluence score.

2. Logic Layering & Confluence¶

The script’s intelligence resides in its hierarchical filtering and confluence-based scoring system, designed to maximize the signal-to-noise ratio.

• Interaction Dynamics: The engine primarily seeks Confluence and Mean Reversion within a Trend.

  • Core Signal: A crossover of the TSI and its signal line.

  • Primary Setup (Mean Reversion): The highest-value signal is a crossover occurring within an extreme zone (e.g., bullish crossover in the oversold zone). This represents the end of a pullback and the potential resumption of the primary trend.

  • Secondary Setup (Momentum): A signal is also generated when the TSI exits an extreme zone (e.g., crosses back above the oversold line). This is a momentum-based trigger.

  • Divergences: The script actively scans for classic bullish and bearish divergences between price pivots and TSI pivots, which contribute significantly to the confluence score.

• Hierarchical Filtering: Signals must pass through a series of logical gates before being considered valid. The order of operations is critical:

  1. Session Gate: Signals can be entirely blocked or have their potential score penalized based on the active trading session.

  2. Cooldown Gate: A dynamic cooldown period, adjusted for both volatility (ATR) and consolidation (BB Width), prevents signal clustering.

  3. Slope Gate: Crossovers with insufficient momentum (a shallow angle, measured by crossSlope) are immediately filtered out.

  4. Regime Gate (Anti-Whipsaw): If the BB Width module detects consolidation (isConsolidation), the signal’s score is multiplied by a penalty factor (consolMultiplier), and the cooldown is extended. This acts as a powerful filter against chop.

  5. Trend Gate (EMA + ADX): For signals not occurring in extreme zones, they are gated by the primary trend filters. A bullish signal is only considered if close > EMA and ADX > Threshold. This forces the script to trade pullbacks, not reversals against a strong trend.

  6. Multi-Timeframe Gate: An optional but important filter where the higher-timeframe TSI must either be aligned with the signal direction or, at a minimum, not be in the opposite extreme zone.

• Confluence Scoring: Instead of a binary true/false outcome, the script quantifies signal quality via a weighted scoring system (rawScoreBull/rawScoreBear). Points are awarded for:

  • Signal Type: Base score for a crossover or momentum signal.

  • Location: Bonus for signals in OB/OS zones.

  • Confirmation: Bonuses for ADX confirmation, MTF alignment, structural alignment (BOS/CHoCH), divergence, and strong crossover slope.

  • Volume: A bonus is added for high relative volume; a penalty is subtracted for low volume.

  • Final Adjustment: The raw score is then multiplied by factors for session, composite bias, and consolidation state to produce the final adjScore. A signal is only generated if this score surpasses a user-defined threshold.

3. The Execution Engine¶

The final trigger and risk management logic are defined by a combination of boolean conditions and mathematical constants.

• Boolean Logic (Trigger): A bullish signal (bullSignal) is true if the cooldown period is over (bullCDOK) AND one of the following is true:

  1. Crossover Signal (bullSignalCross):

     • A crossover occurs (ta.crossover(tsi, sig)).

     • It passes the slope filter (slopeOK).

     • It passes the session filter (sessionAllowSignal).

     • AND it meets one of these path-dependent conditions:

       • Path A (Extreme Zone): The crossover occurs within the oversold zone (inOSZone). This is the highest-probability path and bypasses the main trend filter.

       • Path B (Near Extreme Zone): The crossover occurs near the oversold zone (nearOSZone) AND the ADX confirms a strong trend (adxOK).

       • Path C (Standard Trend): The crossover occurs outside extreme zones BUT aligns with all primary filters: trend EMA (trendBullOK), ADX (adxOK), and MTF (bullMTFOK).

  2. Momentum Signal (bullSignalMom):

     • The TSI crosses back out of the oversold zone (momBullSignal).

     • It passes the session filter (sessionAllowSignal).

• Mathematical Constants & Risk Profile:

  • slATRMult (Default: 1.2): This constant sets the base Stop Loss distance as a multiple of the 14-period ATR. A value of 1.2 implies a relatively tight, volatility-adjusted stop suitable for scalping.

  • tp1RR (Default: 1.5) / tp2RR (Default: 2.5): These constants define the Risk-to-Reward profile. A tp1RR of 1.5 sets the first take-profit target at a distance equal to 1.5 times the initial risk (the SL distance). This establishes a positive reward asymmetry from the outset.

  • Score-Adjusted Stop Loss (calcSLMultiplier): This is a critical risk management function. It modifies the slATRMult based on the signal’s final confluence score.

    • Logic: 1.35 - score * 0.05.

    • Effect: A high-score signal (e.g., 9/10) results in a smaller multiplier (0.9), tightening the stop loss because confidence is high. A low-score signal (e.g., 4/10) results in a larger multiplier (1.15), widening the stop to allow for more noise. This dynamically links the assumed risk of a trade to its quantified quality.

  • consolMultiplier (Default: 0.75): When consolidation is detected, the final score is reduced by 25%. This constant directly controls the script’s sensitivity reduction during ranging markets.

  • alertScoreMin (Default: 4.0): This hard-coded number acts as the final gatekeeper for execution. No matter how many filters a signal passes, if its final confluence score is below this threshold, it is considered low-probability and will not generate an alert or display TP/SL zones.