Quantum Reservoir Computing

21 paź

⚛ Quantum Reservoir Computing - Multi-Scale Market Analysis

OVERVIEW
This indicator combines three structural analysis kernels (Energy, Resonance, Topology) with a 6-spin reservoir computing network to provide multi-dimensional market state monitoring. It is designed to detect structural shifts, coherence alignment, and potential entry timing through visual analytics and optional signal markers.

WHAT MAKES IT ORIGINAL
Unlike single-indicator approaches, QRC fuses complementary analysis methods and uses a reservoir computing layer (coupled oscillator network) to capture temporal market structure. The system uses entropy-compensated signal logic to maintain directional alignment across kernels with inverted mathematical properties.

HOW IT WORKS (Technical Details)

1. ENERGY KERNEL
Measures compression state through two components:
• Entropy: Volatility-normalized return distribution, inverted (low volatility = high compression energy)
• ATR Compression: Short-period ATR divided by longer-period baseline ATR
• Final Energy: Weighted average of both components, ranging 0 to 1

2. RESONANCE KERNEL
Calculates cross-timeframe coherence using:
• 6 exponential moving averages (periods: 9, 14, 20, 30, 48, 84)
• Slope calculation for each EMA
• Amplitude weighting based on user-selected mode (Close/ATR/StDev)
• Coherence Index (CI): Measures directional agreement across all timeframes
• Mode Persistence: Stability of CI over 20 bars

3. TOPOLOGY KERNEL
Analyzes path geometry through:
• Turn density: Rate of price directional changes
• Curvature: Second-order price differences normalized by ATR
• Combined into a 0-1 topology change metric

4. RESERVOIR COMPUTING (6-Spin Network)
Six coupled state variables (spins) arranged in a ring topology:
• Drive signal combines directional consensus, price z-score, volume, and ATR regime
• Each spin updates via hyperbolic tangent activation with neighbor coupling
• Psi (Ψ): Coherence measure (average pairwise spin correlation)
• Spin Direction: Signed average of all spins
• Pulse detection: Positive changes in Ψ, z-scored to detect energy releases

5. FUSION & SCORING
• Magnitude: Weighted combination of all kernels (0 to 1 scale)
• Direction: Blend of EMA slope consensus, basis slope, and spin direction (-1 to 1)
• ScoreSigned: Direction multiplied by Magnitude (drives visuals)
• GateScore: Amplified score used only for signal threshold checks
• Heat: Entanglement measure combining Ψ, CI, and Magnitude

SIGNAL LOGIC (Important: Entropy-Compensated Inversion)
Because the entropy kernel naturally inverts (low volatility = bullish compression), signal logic compensates to maintain directional alignment:
• LONG signals fire when GateScore crosses below the short threshold (bearish GateScore + bullish structure)
• SHORT signals fire when GateScore crosses above the long threshold (bullish GateScore + bearish structure)

This inversion has been visually validated through metric plotting and maintains correct alignment with Resonance and Topology kernels.

Signal gates require:
• Two-of-three pass: CI ≥ minimum, Mode Persistence ≥ minimum, Ψ ≥ minimum
• Heat ≥ minimum threshold
• OR recent pulse window active (ΔΨ edge within N bars)
• Minimum bar spacing between signals (prevents clustering)

VISUAL COMPONENTS

1. Contained Ribbon (Recommended Mode)
• Center line: Basis EMA
• Edge: Positioned by ScoreSigned value
• Fill color: Green (bullish) or Red (bearish)
• Width: ATR-adaptive with configurable floor/ceiling

2. Quantum Aurora (Multi-Layer Energy Bands)
• 5-8 harmonic layers with phase-driven oscillations
• Colors shift with Heat level (cool blue at low Heat, warm orange/magenta at high Heat)
• Creates visual texture that reflects market state dynamics

3. Interference Mesh
• Subtle oscillating overlay modulated by CI and ScoreSigned
• Provides depth perception without visual clutter

4. Resonance Cloud
• Width proportional to Coherence Index
• Wide cloud = strong cross-timeframe alignment
• Narrow cloud = weak structural coherence

5. Energy Particles
• Floating micro-dots with density mapped to Magnitude
• Color-coded by Heat level (gold/cyan/gray)
• Provides continuous conviction feedback

6. Regime Atmosphere
• Background tint indicating market mode:
- Green: Coherent trend (CI>0.65, Ψ>0.55)
- Red: Choppy regime (CI<0.45, Ψ<0.40)
- Purple: Transition state

DASHBOARDS

1. Main Dashboard (Moveable, Resizable)
• Regime indicator with color-coded status
• Horizontal meter gauges for Ψ, CI, Heat, Magnitude
• Signal strength bars for Score and Gate
• Status indicators (dots) for ΔΨ, Heat, CI health
• Directional arrows and bars-since-signal counter
• Size options: Tiny, Small, Normal, Large
• Position: All four corners available

2. Heat HUD (Entanglement Matrix)
• Multi-row gradient display of last N bars (configurable 10-120)
• Metrics: Heat, Psi, CI, Magnitude, Pulse Z-score, Gate proximity
• Color-coded blocks show metric intensity over time
• Live footer with current values
• Resizable and moveable

HOW TO USE

Step 1: Monitor Regime and Structure
• Check Dashboard regime indicator (Trend/Chop/Transition)
• Observe Aurora flow (smooth = stable, erratic = unstable)
• Wide Resonance Cloud indicates strong multi-timeframe alignment

Step 2: Watch Entanglement Heat
• Heat HUD shows persistent structure as amber/red runs
• Green status dots indicate healthy metrics
• Rising Heat + rising Ψ suggests mode-locking

Step 3: Confirm Gate Conditions
• Dashboard displays effective thresholds (dynamically relaxed after dry periods)
• Two-of-three gate (CI/ModePersistence/Ψ) must pass OR recent pulse active
• Strength bars show conviction level

Step 4: Interpret Signals
• Enable "Show Diagnostic Plots" to verify metric behavior on your symbols
• Signals appear as tiny triangles (green below bars = long, red above = short)
• Best confluence: Heat rising + fresh pulse cluster + strong CI

Step 5: Risk Management
• Place stops beyond opposite ribbon edge plus 0.5 ATR buffer
• Trail stops following basis ± ATR fraction while Heat/Psi remain elevated
• Exit early if CI or Ψ collapse (status dots turn yellow/red)

CUSTOMIZATION
Extensive settings available:
• Core: EMA length, ATR length, pulse thresholds, heat minimum
• Signals: Mode (Aggressive/Normal/Conservative), thresholds, spacing, gain
• Visuals: Ribbon mode, Aurora layers, particle density, all show/hide toggles
• Dashboards: Size, position for both main dashboard and heat HUD
• Diagnostics: Optional metric plots for validation

IMPORTANT DISCLAIMERS
• This indicator does not predict future price movements
• Signals use entropy-compensated inversion (explained above); verify on your symbols
• Always backtest on your specific markets and timeframes before live trading
• Past performance does not guarantee future results
• Heavy visuals may impact performance on lower-end devices (use Performance toggles)
• Designed for liquid markets (major indices, forex, crypto); may underperform on illiquid symbols
• Complex system with learning curve; read full guide embedded in code

DIAGNOSTIC MODE
Enable "Show Diagnostic Plots" in settings to verify metric behavior:
• Heat, Psi, CI, Magnitude plotted in lower pane
• ScoreSigned and GateScore normalized to 0-1 scale
• Reference lines at 0.25, 0.5, 0.75 for threshold context
• Observe metric alignment with price action on YOUR symbols

METHODOLOGY NOTE
The "Quantum" terminology refers to the reservoir computing methodology (coupled oscillator network), not actual quantum mechanics. The 6-spin network uses hyperbolic tangent activation functions to model temporal market structure. This is a deterministic mathematical model, not a quantum computing system.

BEST SUITED FOR
• Liquid markets: Major indices (ES, NQ), forex majors (EUR/USD, GBP/USD), large-cap crypto (BTC, ETH)
• Timeframes: 5-minute through daily (works on all, but designed for intraday to swing)
• Trading styles: Structure-based entries, multi-timeframe confluence, visual state monitoring
• Experience level: Intermediate to advanced (complex system with learning curve)

PERFORMANCE CONSIDERATIONS
• Heavy calculations (6 spins, 6 EMAs, Aurora layers, particles) may lag on lower-end devices
• Use "Dashboard Size: Tiny" and reduce "Aurora Layers" to 2-3 for better performance
• Consider disabling "Energy Particles" on mobile devices
• Script is optimized with array capping and label recycling, but complexity remains high

SUPPORT & UPDATES
• Questions about usage or settings? Send me a message - I respond within 24 hours
• Feature requests are welcome for consideration in future updates
• Bug reports appreciated and addressed promptly
• Script will be maintained and updated as needed

FINAL REMINDERS
• This is an analytical tool, not a trading system
• Always backtest on YOUR symbols and timeframes before live use
• Use proper risk management - stops, position sizing, etc.
• Past performance does not guarantee future results
• Start with demo/paper trading to learn the system

— Dskyz, Trade with insight. Trade with anticipation.

9 lis

Informacje o Wersji

Update: Machine Learning Integration
🧠 Core Architecture Overhaul
Replaced 6-Spin Quantum Reservoir with Echo State Network (ESN)

Implemented full reservoir computing framework with configurable network size (10-50 nodes)
Added sparse recurrent matrix with spectral radius control for temporal memory
Integrated online learning via Recursive Least Squares (RLS) algorithm with adaptive forgetting
Dynamic prediction horizon (1-10 bars ahead) with multi-input feature processing
🎯 New ML-Based Signal System
Predictive Signals Replace Pulse Detection

Signals now fire based on ML forecast direction + prediction strength + confidence threshold
Removed pulse-based triggering (ΔΨ Z-score gates); replaced with adaptive ML gating
Added trend filter option (EMA20/50 crossover alignment)
Implemented multi-metric confirmation system (Heat, CI, Magnitude, Psi aggregate scoring)
📊 Enhanced Metrics & Intelligence
New ML Performance Tracking

ML Confidence: Directional accuracy + error-weighted scoring + learning progress
Prediction Strength: Forecast magnitude normalized by volatility
RMSE Monitoring: Real-time prediction error tracking
Directional Accuracy: Running win rate percentage displayed in dashboard
🎨 Dashboard & Visualization Updates
Expanded Information Display

New ML Engine status row (🟢 OPTIMAL / 🟡 TRAINED / 🔴 LEARNING)
ML Confidence gauge with percentage display
ML Forecast direction + strength bars
Confirmation score meter
Real-time accuracy % and RMSE (‰) in status footer
HUD matrix now includes ML Confidence and ML Prediction heat columns
⚙️ Parameter Changes
Removed (legacy pulse system):

Reservoir Pulse k (ΔΨ Z)
Bars After Pulse Allowed
Min Entanglement Heat
Long/Short Thresholds
Signal Score Gain
Added (ML configuration):

Reservoir Nodes (network size)
Spectral Radius (memory depth)
Input Scaling, Leak Rate, Sparsity
Learning Rate, Forget Factor
Warmup Bars, Prediction Horizon
ML Signal Strength threshold
Min ML Confidence threshold
Use ML Gating toggle
Use Trend Filter toggle
Require Multi-Metric Confirmation toggle
🔬 Technical Improvements
Adaptive prediction scaling for robust cross-market performance
Error history buffering for statistically meaningful RMSE
Warmup period prevents premature signals during model training
Direction-weighted ML boost applied to GateScore
More conservative default signal spacing (10→8 bars) with smarter gating
📈 Use Case
Best for traders seeking predictive edge rather than reactive confirmation. The ML engine learns market-specific patterns during warmup, then provides probabilistic forecasts with quantified confidence. Use ML Confidence and Directional Accuracy metrics to assess model reliability before taking signals.