Indicators Description

1. Component Deconstruction¶

The script’s primary analytical component is not a standard library indicator (like RSI or MACD) but a custom-built structural pattern recognition engine. Its core building block is the ta.pivothigh and ta.pivotlow function.

• Pivot Points (ta.pivothigh / ta.pivotlow)

  • Specific Configuration:

    • Source: high for ta.pivothigh and low for ta.pivotlow.

    • Left Strength (Lookback): INPUT_PIVOT_LEFT (default: 5 bars). This defines the number of bars to the left of a potential pivot that must have lower highs (for a pivot high) or higher lows (for a pivot low).

    • Right Strength (Offset): INPUT_PIVOT_RIGHT (default: 5 bars). This defines the number of bars to the right that must have lower highs or higher lows. This introduces a lag of INPUT_PIVOT_RIGHT bars, as a pivot can only be confirmed after this period has passed.

  • Functional Modification: The script does not merely plot these pivots. It transforms them into a structured dataset for advanced analysis.

    1. Data Capture: Upon detection, the price and bar index (bar_index - INPUT_PIVOT_RIGHT) of each pivot are captured into a custom Coordinate object.

    2. Stateful Buffering: These Coordinate objects are stored in two persistent arrays, pivot_highs and pivot_lows. These arrays act as rolling buffers, maintaining a history of the last 20 significant market structure points. This transforms the stateless output of ta.pivothigh/low into a stateful dataset that forms the basis for trendline construction.

2. Logic Layering & Confluence¶

The script employs a hierarchical filtering system to identify high-probability wedge patterns and distinguish them from random market noise. The logic progresses from raw data collection to geometric validation and finally to breakout confirmation.

• Interaction Dynamics: Geometric Confluence The engine’s core function is to detect a specific geometric confluence between two trendlines derived from the pivot point buffers.

  1. Trendline Construction: A potential pattern is considered only when both pivot arrays contain at least MIN_TOUCHES_PER_LINE (default: 3) points. The script then constructs two Trendline objects:

     • Upper Trendline: Connects the N-th most recent pivot high to the most recent pivot high (where N is MIN_TOUCHES_PER_LINE).

     • Lower Trendline: Connects the N-th most recent pivot low to the most recent pivot low.

  2. Slope-Based Classification: The script calculates the slope of each trendline and looks for a specific mathematical relationship to classify the pattern:

     • Rising Wedge (Bearish Implication): Identified if tl_up.slope > 0 AND tl_lo.slope > 0 AND tl_lo.slope > tl_up.slope. This mathematical condition confirms that both lines are rising, but the lower line (support) is rising faster than the upper line (resistance), forcing a convergence.

     • Falling Wedge (Bullish Implication): Identified if tl_up.slope < 0 AND tl_lo.slope < 0 AND tl_up.slope < tl_lo.slope. This confirms both lines are falling, but the upper line (resistance) is falling faster than the lower line (support), again forcing a convergence.

• Hierarchical Filtering (Gatekeeper Logic) The script applies a series of stringent checks to validate a potential pattern before it becomes “active.”

  1. Gatekeeper 1: Data Sufficiency: The entire detection logic is bypassed unless array.size(pivot_highs) and array.size(pivot_lows) are both greater than or equal to MIN_TOUCHES_PER_LINE. This ensures the pattern is based on a statistically relevant number of structural points.

  2. Gatekeeper 2: Geometric Convergence: The slope conditions described above must be met. Any other slope combination (e.g., a channel, a broadening wedge) is ignored.

  3. Gatekeeper 3: Future Apex Projection: The script calculates the theoretical intersection point (the “apex”) of the two trendlines. A pattern is only considered valid if this apex occurs at a future bar index (apex_idx > bar_index). This filters out patterns that have already completed their consolidation phase.

  4. Gatekeeper 4: Pattern Integrity Check: The check_violation() method acts as a critical noise filter. It iterates through all historical bars within the pattern’s boundaries (from its starting pivot to the current bar) and invalidates the entire pattern if any bar’s close has prematurely breached either the upper or lower trendline. This ensures the identified consolidation has remained intact.

3. The Execution Engine¶

The execution engine transitions the script from pattern observation to trade signal generation. It is governed by a precise set of boolean conditions and configurable risk parameters.

• Boolean Logic: The Trigger Condition A trade signal is generated only when a confluence of three conditions is met on the same bar for an “active” wedge pattern:

  1. Directional Breakout: The close of the current bar must cross the pattern boundary in the expected direction.

     • Long Trigger: close > upper_trendline_price for a Falling Wedge.

     • Short Trigger: close < lower_trendline_price for a Rising Wedge.

  2. Momentum Conviction: The body of the breakout candle must represent a significant percentage of the candle’s total range.

     • Formula: bodyPct = (math.abs(close - open) / (high - low)) * 100.0

     • Condition: bodyPct >= MIN_BODY_PCT (default: 50%). This filter is designed to reject weak breakouts or “doji-like” candles that signify indecision rather than conviction.

  3. Invalidation Avoidance: The breakout must not be an “invalidation” (e.g., a close above a rising wedge or below a falling wedge).

  The final signal (longSignal or shortSignal) is true only when correctBreakout AND strongBody evaluates to true. A breakout in the correct direction but with a weak body (correctBreakout AND NOT strongBody) marks the pattern as is_failed, preventing future signals from it.

• Mathematical Constants & Risk Profile Once a valid signal is triggered, the script calculates Entry, Stop Loss (SL), and Take Profit (TP) levels. The risk-to-reward profile is directly influenced by these user-defined settings.

  • Entry Price: close of the breakout candle.

  • Stop Loss (slPrice): The calculation is determined by the SL_MODE input:

    • "Breakout Candle": Uses the low (for longs) or high (for shorts) of the signal candle. This creates a tight, dynamic stop based on recent volatility.

    • "Previous Swing": Uses the price of the most recently confirmed pivot point (recentSwingLow or recentSwingHigh). This provides a wider, structure-based stop.

    • "Fixed Points": entryPrice - (FIXED_SL_POINTS * syminfo.mintick). This uses a static risk distance defined by the user, where FIXED_SL_POINTS acts as a multiplier for the instrument’s minimum price change (syminfo.mintick).

  • Take Profit (tpPrice): The calculation is determined by the TP_MODE input:

    • "Risk Reward": The TP is a direct function of the risk taken.

      • Formula: tpPrice = entryPrice + (riskDistance * RR_RATIO) for a long, where riskDistance = entryPrice - slPrice.

      • The RR_RATIO (default: 1.0) is the key mathematical constant that defines the trade’s reward potential relative to its risk.

    • "Fixed Points": entryPrice + (FIXED_TP_POINTS * syminfo.mintick). This sets a static profit target, independent of the calculated risk.