Indicators Description

1. Component Deconstruction¶

This section deconstructs the individual mathematical and logical components that form the foundation of the RSI Elite Toolkit.

Relative Strength Index (RSI)¶

• Specific Configuration:

  • Price Source: input.source(close, "RSI Source") - Defaults to the bar’s closing price.

  • Length: input.int(21, "RSI Period") - A lookback period of 21 bars, slightly longer than the common 14-period default, aiming to smooth the oscillator and focus on more established momentum shifts.

• Functional Modification:

  • Custom Function: The RSI is calculated within a dedicated function calc_rsi(). This is a standard implementation using ta.rma (Wilder’s Moving Average) for smoothing the up and down components, which is the correct mathematical definition of the RSI. It is not a “hacked” version but rather a clean encapsulation.

  • Non-Repainting MTF Engine: The script’s primary modification is its Multi-Timeframe (MTF) capability. When enable_mtf is true, it uses request.security() to fetch the RSI value from a higher timeframe (mtf_timeframe). Crucially, it employs barmerge.lookahead_off to prevent the engine from using future data, ensuring the signals are non-repainting and historically accurate.

RSI Signal Smoothing (Moving Average of RSI)¶

• Specific Configuration:

  • Type: input.string("SMA", "Smoothing Type", options = ["SMA", "EMA", "WMA", "RMA"]) - A user-selectable moving average type, defaulting to a Simple Moving Average (SMA).

  • Length: input.int(14, "Smoothing Period") - A 14-period lookback.

• Functional Modification:

  • This is not a moving average of price, but a moving average of the rsi_val itself. Its purpose is to act as a “signal line” for the RSI oscillator. The mathematical effect is a second layer of smoothing applied to the momentum calculation. Crossovers between the RSI and its own moving average are used to generate signals with reduced lag compared to price crossing a moving average, but with a better signal-to-noise ratio than raw RSI threshold crosses.

Institutional Bias Filter (Exponential Moving Average)¶

• Specific Configuration:

  • Price Source: Hard-coded to close.

  • Length: input.int(200, "Bias Period") - A fixed 200-period EMA.

• Functional Modification:

  • This is a standard 200 EMA applied to the main chart’s price. It is not mathematically modified. Its function is purely contextual, serving as a binary filter to define the macro market structure (Bullish if close > 200 EMA, Bearish if close < 200 EMA).

Volatility Matrix (Average True Range - ATR)¶

• Specific Configuration:

  • Length: input.int(14, "Volatility Matrix (ATR)") - A standard 14-period ATR.

• Functional Modification:

  • This is a standard ATR calculation. It is not used as a signal generator but as a core component of the risk management engine. Its value provides a dynamic, volatility-adjusted unit of measurement for calculating Stop Loss and Take Profit distances.

Divergence Intelligence Engine¶

• Specific Configuration:

  • Pivot Lookback: div_pivot_left (5 bars), div_pivot_right (5 bars) - Defines the window for identifying a pivot high/low on the RSI. A pivot low, for example, is a point where the RSI is lower than the 5 bars before it and the 5 bars after it.

  • Analysis Range: div_max_range (60 bars), div_min_range (5 bars) - Constrains the search for a previous pivot. A divergence is only considered valid if the distance between the two pivots is between 5 and 60 bars.

• Functional Modification:

  • This is a custom-built logical engine, not a standard indicator. Its mechanics are as follows:

    1. Pivot Identification: It uses ta.pivotlow() and ta.pivothigh() on the rsi_val to locate significant turning points in momentum.

    2. State Capture: Upon finding a pivot, ta.valuewhen() is used to capture the RSI value and the corresponding low or high price at that exact bar. It stores both the most recent pivot ([0]) and the one prior ([1]).

    3. Distance Validation: ta.barssince() calculates the number of bars between the two pivots. This distance is checked against div_min_range and div_max_range to filter out divergences that are either too close together (noise) or too far apart (irrelevant).

    4. Logical Matrix: It applies the classic definitions of divergence:

       • Regular Bullish: price_pl < price_pl_prev (Lower Low in Price) AND rsi_pl > rsi_pl_prev (Higher Low in RSI).

       • Regular Bearish: price_ph > price_ph_prev (Higher High in Price) AND rsi_ph < rsi_ph_prev (Lower High in RSI).

2. Logic Layering & Confluence¶

The script’s primary innovation is its departure from single-trigger logic in favor of a weighted, score-based confluence engine. It does not use a strict hierarchical filter but rather assigns evidentiary weight to different market phenomena.

• Interaction Dynamics: The engine synthesizes signals from three distinct types of events:

  1. Threshold Crosses: The RSI crossing above the level_os (Oversold) or below the level_ob (Overbought) threshold. This is the most basic form of mean-reversion signal.

  2. Crossovers: The rsi_val crossing its own smoothing_ma. This is a momentum-shift signal.

  3. Divergences: The sophisticated price-momentum decoupling identified by the Divergence Engine. This is a momentum-exhaustion signal.

• Hierarchical Filtering (via Weighted Scoring): The script achieves filtering not by a sequence of gates, but by assigning points to each condition. This creates a flexible hierarchy where a high-value event can trigger a signal on its own, while lower-value events must occur in combination.

  • Divergence (3 points): Given the highest weight, as it represents a significant structural failure in momentum. A confirmed divergence is the most powerful single component.

  • RSI Extreme Exit (2 points): A strong confirmation that short-term pressure is abating.

  • Trend Bias (1 point): The 200 EMA acts as a contextual modifier. A trade aligned with the macro trend is considered higher probability and is awarded a point.

  • RSI MA Cross (1 point): The lowest weight, representing an early, and potentially noisy, shift in momentum.

A signal is generated only when the cumulative long_score or short_score meets or exceeds the min_confluence threshold (defaulting to 3). This means a signal can be triggered by Divergence alone (3 pts), or by RSI Extreme Exit + Trend Bias (2+1=3 pts), or by RSI Extreme Exit + RSI MA Cross (2+1=3 pts).

3. The Execution Engine¶

Trigger Conditions¶

• Boolean Logic: The final trigger is a simple Boolean comparison:

  • buy_signal = long_score >= min_confluence

  • sell_signal = short_score >= min_confluence

• The script also includes a state machine (var int active_state) which prevents new signals from being generated while a trade visualization is already active on the chart (active_state != 0). A new signal can only trigger if active_state == 0.

Exit Conditions & Mathematical Constants¶

The script does not generate dynamic exit signals. Instead, it projects a static Stop Loss (SL) and Take Profit (TP) zone based on volatility at the time of the trigger.

• Mathematical Constants:

  • sl_multiplier: A float (default 3.0) that scales the ATR value to determine the stop loss distance.

  • tp_multiplier: A float (default 8.0) that scales the ATR value to determine the take profit distance.

• Risk-to-Reward Profile: The script’s design hard-codes a specific risk-to-reward ratio based on the input multipliers.

  • Stop Loss Calculation:

    • Long: entry_px - (ta.atr(atr_period) * sl_multiplier)

    • Short: entry_px + (ta.atr(atr_period) * sl_multiplier)

  • Take Profit Calculation:

    • Long: entry_px + (ta.atr(atr_period) * tp_multiplier)

    • Short: entry_px - (ta.atr(atr_period) * tp_multiplier)

  • Implied R:R: With default settings, the script visualizes a trade with a fixed Risk:Reward ratio of tp_multiplier / sl_multiplier = 8.0 / 3.0 ≈ 2.67:1. The use of ATR makes this calculation dynamic to market volatility; the stop and target are wider in volatile markets and tighter in calm markets, but their ratio remains constant.

• Trade Termination (Visualization Only): The active_state is reset to 0 (allowing for a new signal) if price touches the calculated SL or TP levels.

  • Long Exit: low <= stop_level or high >= target_level

  • Short Exit: high >= stop_level or low <= target_level This logic governs the termination of the visual boxes on the chart, it does not represent a backtested strategy exit.