FinRobot - Multi-Strategy Algorithmic Trading System

Status: Active Development
Last Updated: 2026-05-03
Version: 2.0 - Multi-Strategy Engine

---

Overview

FinRobot is a self-improving autonomous algorithmic trading system with a closed feedback loop. It combines multiple trading strategies including Grid Trading, Smart Money Concepts (SMC), Harmonic Patterns, and portfolio-level risk management with continuous genetic optimization.

Key Features

• Multi-Strategy Ensemble: Combines Grid, SMC, Harmonic strategies
• Smart Money Concepts: Order Blocks, Fair Value Gaps, Liquidity Sweeps
• Harmonic Patterns: Gartley, Bat, Butterfly, Crab, Shark, Cypher patterns
• Genetic Optimization: Continuous parameter optimization using genetic algorithms
• Portfolio Management: Dynamic capital allocation with risk management
• Continuous Learning: Self-improving through iterative backtesting

---

Project Structure

FinRobot/
├── finrobot/
│   ├── smart_money_concepts.py    # SMC module (Order Blocks, FVG, Sweeps)
│   ├── harmonic_patterns.py       # Harmonic pattern detection
│   ├── genetic_optimizer.py       # Genetic algorithm optimizer
│   ├── portfolio_manager.py       # Portfolio allocation & risk management
│   └── multi_strategy_engine.py   # Main engine combining all strategies
├── multi_strategy_runner.py       # CLI runner for continuous optimization
├── MULTI_STRATEGY_PORTFOLIO_PLAN.md  # Master plan document
├── AGENTS.md                      # Agent documentation
└── README.md                      # This file

---

Architecture

Core Components

1. Smart Money Concepts (smart_money_concepts.py)

• Order Blocks: Bullish/Bearish OB detection with strength scoring
• Fair Value Gaps: Imbalance detection between price levels
• Liquidity Sweeps: Stop hunt detection and reversal signals
• Signal Generation: Combined setup detection for high-probability trades

2. Harmonic Patterns (harmonic_patterns.py)

• Pattern Detection: 6 classic harmonic patterns
  • Gartley (0.618, 1.27, 0.786 ratios)
  • Bat (0.382-0.5, 1.618-2.618, 0.886)
  • Butterfly (0.786, 1.618-2.24, 1.27)
  • Crab (0.382-0.618, 2.24-3.618, 1.618)
  • Shark (1.13-1.618, 1.618-2.24, 0.886-1.13)
  • Cypher (0.382-0.618, 1.13-1.414, 0.786)
• Swing Point Detection: Automatic significant high/low identification
• Fibonacci Validation: Ratio matching with tolerance
• Signal Generation: Entry, SL, TP levels with confidence scoring

3. Genetic Optimizer (genetic_optimizer.py)

• Population Management: 50 individuals per generation (configurable)
• Selection: Tournament selection (size 3)
• Crossover: Single-point crossover (rate 0.8)
• Mutation: Adaptive mutation (rate 0.15)
• Elitism: Top 5 individuals preserved
• Fitness Function: Multi-objective (win rate, return, drawdown)
• Parameter Types: Supports float, int, and choice parameters

4. Portfolio Manager (portfolio_manager.py)

• Performance Tracking: Per-strategy metrics (Sharpe, drawdown, win rate)
• Allocation Methods:
  • Equal weight
  • Inverse volatility (risk parity)
  • Sharpe-weighted
  • Momentum-based
• Rebalancing: Time-based (hourly, daily, weekly) or drift-based (>5%)
• Risk Management:
  • Max drawdown limits
  • Daily/weekly loss limits
  • Correlation monitoring
• Circuit Breakers: Auto-stop on excessive drawdown

5. Multi-Strategy Engine (multi_strategy_engine.py)

• Strategy Implementations:
  • GridStrategy: Grid trading with trend filter
  • SMCStrategy: Smart Money Concepts with OB + Sweep setups
  • HarmonicStrategy: Harmonic pattern trading
• Signal Structure: Unified format across all strategies
• Backtest Engine: Event-driven simulation
• Optimization Loop: Periodic genetic algorithm runs
• State Management: Save/load full system state

---

Usage

Starting the Optimizer

# Start fresh optimization
python multi_strategy_runner.py --iterations 10000 --capital 10000

# Resume from saved state
python multi_strategy_runner.py --resume

# Check current status
python multi_strategy_runner.py --status

Monitoring Progress

# Watch logs in real-time
tail -f multi_strategy.log

# Check status anytime
python multi_strategy_runner.py --status

Python API

from multi_strategy_runner import ContinuousOptimizer

# Create optimizer
optimizer = ContinuousOptimizer(
    initial_capital=10000.0,
    iterations=10000
)

# Start optimization
optimizer.start()

# Or run single iteration
optimizer.run_single_iteration()

# Get current status
summary = optimizer.get_summary()
print(f"Win Rate: {summary['best_win_rate']:.2%}")

---

Configuration

Strategy Parameters

Each strategy has configurable parameters optimized by genetic algorithm:

Grid Strategy:

• grid_step_pips: Distance between grid levels (0.5 - 5.0)
• take_profit_pips: TP distance (1.0 - 10.0)
• max_grid_levels: Max open positions (2 - 10)
• trend_ema_fast/slow: Trend filter periods
• volatility_filter: ATR-based volatility threshold

SMC Strategy:

• ob_lookback: Periods for OB detection (3 - 10)
• fvg_min_size: Minimum FVG size (0.0005 - 0.005)
• min_ob_strength: Minimum OB confidence (0.5 - 0.9)
• min_sweep_strength: Minimum sweep confidence (0.6 - 0.9)
• risk_reward_min: Minimum R:R ratio (1.0 - 3.0)

Harmonic Strategy:

• tolerance: Fibonacci ratio tolerance (0.02 - 0.1)
• min_confidence: Minimum pattern confidence (0.5 - 0.95)
• swing_window: Swing point detection window (3 - 10)
• risk_reward_min: Minimum R:R ratio (1.0 - 3.0)

Genetic Algorithm Settings

GeneticOptimizer(
    population_size=50,      # Individuals per generation
    elite_size=5,            # Top individuals to preserve
    mutation_rate=0.15,      # Mutation probability
    crossover_rate=0.8,      # Crossover probability
    tournament_size=3        # Tournament selection size
)

Optimization Targets

• Win Rate: Target > 65%
• Return: Target > 5% per session
• Max Drawdown: Limit < 3%
• Sharpe Ratio: Target > 2.0

---

Performance Metrics

The system tracks:

• Win Rate: Percentage of winning trades
• Total Return: Cumulative percentage return
• Sharpe Ratio: Risk-adjusted return metric
• Maximum Drawdown: Largest peak-to-trough decline
• Profit Factor: Gross profit / gross loss
• Volatility: Standard deviation of returns
• Expectancy: Average R per trade

---

Development Roadmap

Completed (v2.0)

• ✅ Multi-strategy ensemble architecture
• ✅ Smart Money Concepts implementation
• ✅ Harmonic pattern detection
• ✅ Genetic optimization engine
• ✅ Portfolio management system
• ✅ Continuous optimization loop
• ✅ State persistence (save/load)

In Progress

• 🔄 Live data integration (MT5 API)
• 🔄 Real-time execution engine
• 🔄 Advanced risk management (Kelly criterion)
• 🔄 Machine learning integration

Planned (v3.0)

• 📋 Deep learning price prediction
• 📋 Sentiment analysis integration
• 📋 Options/Futures strategies
• 📋 Multi-asset correlation trading
• 📋 Real-time portfolio optimization
• 📋 Web dashboard for monitoring

---

Troubleshooting

Common Issues

1. Import Errors

# Ensure finrobot is in Python path
import sys
sys.path.insert(0, '/home/openclaw/FinRobot')

2. Memory Issues

• Reduce population_size in GeneticOptimizer
• Decrease iterations in runner
• Clear data cache periodically

3. Slow Optimization

• Reduce generation count per optimization
• Decrease population size
• Use parallel processing (add multiprocessing)

4. No Signals Generated

• Check data quality and format
• Verify parameter ranges are appropriate
• Increase lookback periods for indicators
• Check minimum confidence thresholds

Debug Mode

import logging
logging.getLogger().setLevel(logging.DEBUG)

# Run with verbose output
optimizer = ContinuousOptimizer()
optimizer.run_single_iteration()  # Step through one iteration

Getting Help

1. Check logs: tail -f multi_strategy.log
2. Check status: python multi_strategy_runner.py --status
3. Review state: cat optimizer_state.json | jq .

---

Contributing

When adding new strategies:

1. Create new file in finrobot/ directory
2. Inherit from StrategyBase class
3. Implement generate_signal() method
4. Define get_param_specs() for optimization
5. Register in MultiStrategyEngine._initialize_strategies()
6. Add tests and documentation

Code Style

• Follow PEP 8
• Type hints for function signatures
• Docstrings for all public methods
• Logging instead of print statements

---

License

This project is proprietary and confidential.

---

Contact

For questions or issues:

1. Check this README and AGENTS.md
2. Review logs and state files
3. Consult MULTI_STRATEGY_PORTFOLIO_PLAN.md for architecture details

---

Last Updated: 2026-05-03
System Version: 2.0
Status: Production Ready - Continuous Optimization Active