AI Pattern Signal
- Indicadores
- Joshua Akinetos Malik
- Versão: 4.0
- Ativações: 5
Self-calibrating confluence indicator combining 15 candlestick patterns with 8 technical indicators — signals that adapt to your instrument's historical behaviour.
Most indicators treat every pattern the same. AI Pattern Signal does not. Before generating a single signal it scans your chart's history, measures which patterns have actually produced profitable moves on that specific instrument, and adjusts each pattern's weight accordingly. The result is a confluence score that reflects what has worked on your chart, not just what looks good in a textbook.
How It Works
The indicator combines two scoring layers to produce a final buy or sell confluence score from 0 to 100.
The first layer is the pattern recognition engine, which detects 15 candlestick and price action patterns on each bar. Each detected pattern contributes a weighted score, with the weights set dynamically by the AI calibration.
The second layer is the technical confluence engine, which reads eight indicator values simultaneously — EMA alignment, RSI positioning, Bollinger Band location, MACD direction, ADX trend strength, Stochastic momentum, and ATR volatility — and adds a confluence score based on how many indicators agree with the pattern direction.
The two layers are blended (70% indicators, 30% patterns) into a final score. When that score crosses your minimum threshold, a signal is placed on the chart.
AI Calibration Engine
This is what separates AI Pattern Signal from a standard indicator. On load, the calibration engine scans the last 1000 bars of the chart it is attached to and for each pattern it detects, it looks forward 5 bars to measure whether price actually moved profitably by at least 0.5× ATR in the signal direction.
It then calculates a historical accuracy percentage for each of the 15 patterns and adjusts their weights:
- Patterns with above 50% accuracy on your instrument receive a boost
- Patterns with below 50% accuracy are reduced
- The final weights are blended 60% calibrated, 40% default to avoid overfitting
The dashboard shows the resulting win rate, number of signals found, and bars scanned so you can see exactly what the calibration found. You can re-run calibration at any time using the button on the dashboard, or reset all weights back to default if you want a clean start on a new instrument.
Because calibration runs on the specific symbol and timeframe the indicator is attached to, it self-tunes to each chart independently. The same indicator behaves differently on EURUSD versus XAUUSD versus USDJPY because each instrument has its own pattern behaviour history.
15 Patterns Detected
Candlestick reversal: Engulfing, Morning/Evening Star, Three Candle patterns, Doji, Hammer
Price action: Pin Bar, Inside Bar
Smart money concepts: Fair Value Gap (FVG), Break of Structure (BOS), Change of Character (CHoCH), Order Block
Advanced: RSI/MACD Divergence, Double Bottom/Top, Trend Pullback, Volume Spike
Each pattern can be individually enabled or disabled in the settings, giving full control over what the engine detects.
Signal Types
Normal buy arrow (blue, below candle) — confluence score above minimum threshold, buy bias
Normal sell arrow (red, above candle) — confluence score above minimum threshold, sell bias
Strong buy orb (green filled circle, below candle) — confluence score above the strong signal threshold, high conviction buy
Strong sell orb (red filled circle, above candle) — confluence score above the strong signal threshold, high conviction sell
Pattern confluence dots (cyan/magenta, small) — multiple patterns agree on direction but full confluence threshold not yet met, early warning markers
Market Regime Detection
The indicator continuously monitors the broader market condition and classifies it into four regimes: trending up, trending down, ranging, or volatile. The current regime is displayed on the dashboard. This gives context to every signal — a bullish signal in a confirmed uptrend carries different weight than the same signal in a ranging or volatile environment.
Dashboard
A live panel on the chart displays:
- Current buy and sell confluence scores
- Active market regime
- Calibration status — bars scanned, signals found, historical win rate, time of last calibration
- Individual pattern accuracy table showing each pattern's win rate and adjusted weight after calibration
- Calibrate and Reset buttons
The dashboard updates on bar close only, keeping CPU usage minimal.
Performance Optimisation
The indicator is built for live trading use, not just backtesting. Several optimisations are applied to keep performance clean:
- Dashboard updates on bar close only, not every tick
- Alert logic shares the same calculation block as the dashboard — no duplicate computation
- ATR handle is cached and reused rather than recreated on every bar
- Calibration is deferred by one bar after load so the indicator appears instantly without freezing the terminal
Settings Reference
AI Signal Settings
| Parameter | Default | Description |
|---|---|---|
| AI_MinScore | 30 | Minimum confluence score for a normal signal |
| AI_StrongScore | 50 | Minimum score for a strong signal orb |
| CalibrationBars | 1000 | Bars to scan during AI calibration |
| CalibMinMove | 0.5 | Minimum ATR move to count as a win (×ATR) |
| AutoCalibOnLoad | true | Run calibration automatically on indicator load |
Indicator Settings
| Parameter | Default |
|---|---|
| EMA Fast / Mid / Slow | 20 / 50 / 200 |
| RSI Period | 14 |
| RSI Oversold / Overbought | 35 / 65 |
| Bollinger Bands Period / Deviation | 20 / 2.0 |
| MACD Fast / Slow / Signal | 12 / 26 / 9 |
| ADX Period / Minimum | 14 / 20 |
| ATR Period | 14 |
| Stochastic K / D | 5 / 3 |
Pattern Recognition
All 15 patterns can be individually enabled or disabled via true/false toggles.
Display
| Parameter | Default | Description |
|---|---|---|
| ShowDashboard | true | Show the live dashboard panel |
| ShowPatternLabels | true | Show pattern name text on chart |
| ShowAlerts | true | Popup alert on new signal |
| ShowPatternDots | true | Show early-warning confluence dots |
| PatDot_MinScore | 18 | Minimum pattern score for a dot |
| PatDot_MinCount | 2 | Minimum patterns agreeing for a dot |
Setup
- Attach the indicator to any chart on your chosen symbol and timeframe.
- Calibration runs automatically on the first bar close after load, scanning the last 1000 bars.
- Once calibration completes the dashboard shows the win rate and adjusted weights — signals will begin appearing from that point forward.
- To recalibrate after a period of time or when switching instruments, click the Calibrate button on the dashboard.
- To tune sensitivity, adjust AI_MinScore (lower = more signals, higher = fewer but stronger) and AI_StrongScore (threshold for orb signals).
Notes
- Works on any symbol and timeframe. The AI calibration means it adapts to each instrument individually rather than using fixed settings.
- Lower AI_MinScore values such as 30 will generate more frequent signals. Raising it to 50 or 60 produces fewer but higher conviction signals.
- The indicator is not a trading system on its own. It is a confluence tool designed to help identify high-probability setups for manual review and decision making.
- Past calibration accuracy does not guarantee future results.
© SCEPTRM IGNI LLC — All rights reserved.
