Add strategy management system with multiple trading strategies
- Introduced a new strategies module containing the StrategyManager class to orchestrate multiple trading strategies. - Implemented StrategyBase and StrategySignal as foundational components for strategy development. - Added DefaultStrategy for meta-trend analysis and BBRSStrategy for Bollinger Bands + RSI trading. - Enhanced documentation to provide clear usage examples and configuration guidelines for the new system. - Established a modular architecture to support future strategy additions and improvements.
This commit is contained in:
Vasily.onl
330
docs/strategy_manager.md
Normal file
330
docs/strategy_manager.md
Normal file
@@ -0,0 +1,330 @@
|
||||
# TCP Cycles Strategy Management System
|
||||
|
||||
The Strategy Manager system provides a flexible framework for implementing, combining, and managing multiple trading strategies within the TCP Cycles project.
|
||||
|
||||
## 🏗️ Architecture
|
||||
|
||||
### Module Structure
|
||||
```
|
||||
cycles/
|
||||
├── strategies/ # Strategy management module
|
||||
│ ├── __init__.py # Module exports and version info
|
||||
│ ├── base.py # Base classes (StrategyBase, StrategySignal)
|
||||
│ ├── default_strategy.py # Meta-trend strategy implementation
|
||||
│ ├── bbrs_strategy.py # Bollinger Bands + RSI strategy
|
||||
│ └── manager.py # StrategyManager and orchestration
|
||||
├── Analysis/ # Technical analysis tools
|
||||
├── utils/ # Utility functions
|
||||
├── backtest.py # Backtesting engine
|
||||
└── charts.py # Charting and visualization
|
||||
```
|
||||
|
||||
### Core Components
|
||||
|
||||
1. **`StrategyBase`**: Abstract base class that all strategies inherit from
|
||||
2. **`StrategySignal`**: Represents trading signals with confidence levels and metadata
|
||||
3. **`DefaultStrategy`**: Implementation of the meta-trend strategy using Supertrend indicators
|
||||
4. **`BBRSStrategy`**: Implementation of the Bollinger Bands + RSI strategy with market regime detection
|
||||
5. **`StrategyManager`**: Orchestrates multiple strategies and combines their signals
|
||||
|
||||
### Signal Types
|
||||
- **`"ENTRY"`**: Strategy suggests entering a position
|
||||
- **`"EXIT"`**: Strategy suggests exiting a position
|
||||
- **`"HOLD"`**: Strategy suggests no action
|
||||
|
||||
## 📋 Configuration
|
||||
|
||||
### Single Strategy Configuration
|
||||
|
||||
**Default Strategy Only:**
|
||||
```json
|
||||
{
|
||||
"start_date": "2025-03-01",
|
||||
"stop_date": "2025-03-15",
|
||||
"initial_usd": 10000,
|
||||
"timeframes": ["15min"],
|
||||
"stop_loss_pcts": [0.03, 0.05],
|
||||
"strategies": [
|
||||
{
|
||||
"name": "default",
|
||||
"weight": 1.0,
|
||||
"params": {}
|
||||
}
|
||||
],
|
||||
"combination_rules": {
|
||||
"entry": "any",
|
||||
"exit": "any",
|
||||
"min_confidence": 0.5
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
**BBRS Strategy Only:**
|
||||
```json
|
||||
{
|
||||
"strategies": [
|
||||
{
|
||||
"name": "bbrs",
|
||||
"weight": 1.0,
|
||||
"params": {
|
||||
"bb_width": 0.05,
|
||||
"bb_period": 20,
|
||||
"rsi_period": 14,
|
||||
"strategy_name": "MarketRegimeStrategy"
|
||||
}
|
||||
}
|
||||
]
|
||||
}
|
||||
```
|
||||
|
||||
### Multiple Strategy Configuration
|
||||
|
||||
```json
|
||||
{
|
||||
"strategies": [
|
||||
{
|
||||
"name": "default",
|
||||
"weight": 0.6,
|
||||
"params": {}
|
||||
},
|
||||
{
|
||||
"name": "bbrs",
|
||||
"weight": 0.4,
|
||||
"params": {
|
||||
"bb_width": 0.05,
|
||||
"strategy_name": "MarketRegimeStrategy"
|
||||
}
|
||||
}
|
||||
],
|
||||
"combination_rules": {
|
||||
"entry": "weighted_consensus",
|
||||
"exit": "any",
|
||||
"min_confidence": 0.6
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
## 🔧 Combination Rules
|
||||
|
||||
### Entry Signal Combination Methods
|
||||
|
||||
- **`"any"`**: Enter if ANY strategy signals entry above min_confidence
|
||||
- **`"all"`**: Enter only if ALL strategies signal entry above min_confidence
|
||||
- **`"majority"`**: Enter if more than 50% of strategies signal entry
|
||||
- **`"weighted_consensus"`**: Enter based on weighted average confidence
|
||||
|
||||
### Exit Signal Combination Methods
|
||||
|
||||
- **`"any"`**: Exit if ANY strategy signals exit (recommended for risk management)
|
||||
- **`"all"`**: Exit only if ALL strategies agree on exit
|
||||
- **`"priority"`**: Exit based on priority: STOP_LOSS > SELL_SIGNAL > others
|
||||
|
||||
### Parameters
|
||||
|
||||
- **`min_confidence`**: Minimum confidence threshold (0.0 to 1.0)
|
||||
- **`weight`**: Strategy weight for weighted calculations
|
||||
|
||||
## 🚀 Usage Examples
|
||||
|
||||
### Running with Default Strategy
|
||||
```bash
|
||||
python main.py config_default.json
|
||||
```
|
||||
|
||||
### Running with BBRS Strategy
|
||||
```bash
|
||||
python main.py config_bbrs.json
|
||||
```
|
||||
|
||||
### Running with Combined Strategies
|
||||
```bash
|
||||
python main.py config_combined.json
|
||||
```
|
||||
|
||||
### Running without Config (Interactive)
|
||||
```bash
|
||||
python main.py
|
||||
```
|
||||
|
||||
### Programmatic Usage
|
||||
```python
|
||||
from cycles.strategies import create_strategy_manager
|
||||
|
||||
# Create strategy manager from config
|
||||
config = {
|
||||
"strategies": [
|
||||
{"name": "default", "weight": 0.7, "params": {}},
|
||||
{"name": "bbrs", "weight": 0.3, "params": {"bb_width": 0.05}}
|
||||
],
|
||||
"combination_rules": {
|
||||
"entry": "weighted_consensus",
|
||||
"exit": "any",
|
||||
"min_confidence": 0.6
|
||||
}
|
||||
}
|
||||
|
||||
strategy_manager = create_strategy_manager(config)
|
||||
```
|
||||
|
||||
## ⚙️ Strategy Parameters
|
||||
|
||||
### Default Strategy Parameters
|
||||
- **`stop_loss_pct`**: Stop loss percentage (default: 0.03)
|
||||
|
||||
### BBRS Strategy Parameters
|
||||
- **`bb_width`**: Bollinger Band width (default: 0.05)
|
||||
- **`bb_period`**: Bollinger Band period (default: 20)
|
||||
- **`rsi_period`**: RSI period (default: 14)
|
||||
- **`trending_rsi_threshold`**: RSI thresholds for trending market [low, high]
|
||||
- **`trending_bb_multiplier`**: BB multiplier for trending market
|
||||
- **`sideways_rsi_threshold`**: RSI thresholds for sideways market [low, high]
|
||||
- **`sideways_bb_multiplier`**: BB multiplier for sideways market
|
||||
- **`strategy_name`**: Strategy implementation name
|
||||
- **`SqueezeStrategy`**: Enable squeeze strategy (boolean)
|
||||
- **`stop_loss_pct`**: Stop loss percentage (default: 0.05)
|
||||
|
||||
## 🔌 Adding New Strategies
|
||||
|
||||
### 1. Create Strategy Class
|
||||
|
||||
Create a new file in `cycles/strategies/` (e.g., `my_strategy.py`):
|
||||
|
||||
```python
|
||||
from .base import StrategyBase, StrategySignal
|
||||
|
||||
class MyStrategy(StrategyBase):
|
||||
def __init__(self, weight=1.0, params=None):
|
||||
super().__init__("my_strategy", weight, params)
|
||||
|
||||
def initialize(self, backtester):
|
||||
# Initialize your strategy indicators
|
||||
self.initialized = True
|
||||
|
||||
def get_entry_signal(self, backtester, df_index):
|
||||
# Implement entry logic
|
||||
if entry_condition:
|
||||
return StrategySignal("ENTRY", confidence=0.8)
|
||||
return StrategySignal("HOLD", confidence=0.0)
|
||||
|
||||
def get_exit_signal(self, backtester, df_index):
|
||||
# Implement exit logic
|
||||
if exit_condition:
|
||||
return StrategySignal("EXIT", confidence=1.0,
|
||||
metadata={"type": "MY_EXIT"})
|
||||
return StrategySignal("HOLD", confidence=0.0)
|
||||
```
|
||||
|
||||
### 2. Register Strategy
|
||||
|
||||
Update `cycles/strategies/manager.py` in the `_load_strategies` method:
|
||||
|
||||
```python
|
||||
elif name == "my_strategy":
|
||||
from .my_strategy import MyStrategy
|
||||
strategies.append(MyStrategy(weight, params))
|
||||
```
|
||||
|
||||
### 3. Export Strategy
|
||||
|
||||
Update `cycles/strategies/__init__.py`:
|
||||
|
||||
```python
|
||||
from .my_strategy import MyStrategy
|
||||
|
||||
__all__ = [
|
||||
# ... existing exports ...
|
||||
'MyStrategy'
|
||||
]
|
||||
```
|
||||
|
||||
## 📊 Performance Features
|
||||
|
||||
### Strategy Analysis
|
||||
- Individual strategy performance tracking
|
||||
- Combined strategy performance metrics
|
||||
- Signal quality analysis
|
||||
- Confidence level monitoring
|
||||
|
||||
### Plotting Support
|
||||
- Automatic chart generation for BBRS strategies
|
||||
- Meta-trend visualization for default strategy
|
||||
- Combined signal overlays
|
||||
- Performance comparison charts
|
||||
|
||||
## 🔄 Backward Compatibility
|
||||
|
||||
The system maintains full backward compatibility:
|
||||
- ✅ Existing code using single strategies works unchanged
|
||||
- ✅ Legacy strategy functions are preserved in main.py
|
||||
- ✅ Default behavior matches original implementation
|
||||
- ✅ Gradual migration path available
|
||||
|
||||
## 📚 Best Practices
|
||||
|
||||
### 1. **Risk Management**
|
||||
- Use `"any"` exit rule for faster risk exits
|
||||
- Set appropriate stop loss percentages per strategy
|
||||
- Monitor combined drawdown vs individual strategies
|
||||
|
||||
### 2. **Signal Quality**
|
||||
- Set appropriate `min_confidence` based on strategy reliability
|
||||
- Test individual strategies thoroughly before combining
|
||||
- Monitor signal frequency and quality
|
||||
|
||||
### 3. **Weight Distribution**
|
||||
- Balance strategy weights based on historical performance
|
||||
- Consider strategy correlation when setting weights
|
||||
- Regularly rebalance based on changing market conditions
|
||||
|
||||
### 4. **Testing & Validation**
|
||||
- Backtest individual strategies first
|
||||
- Test combinations on historical data
|
||||
- Validate on out-of-sample data
|
||||
|
||||
### 5. **Monitoring**
|
||||
- Log strategy initialization and errors
|
||||
- Track individual vs combined performance
|
||||
- Monitor signal generation frequency
|
||||
|
||||
## 🔍 Troubleshooting
|
||||
|
||||
### Strategy Not Found Error
|
||||
```
|
||||
ValueError: Unknown strategy: my_strategy
|
||||
```
|
||||
**Solution**: Ensure strategy is registered in `manager.py` `_load_strategies` method
|
||||
|
||||
### No Signals Generated
|
||||
**Possible Causes**:
|
||||
- Strategy initialization failed
|
||||
- Data insufficient for strategy requirements
|
||||
- `min_confidence` threshold too high
|
||||
|
||||
**Solution**: Check logs, verify data, adjust confidence threshold
|
||||
|
||||
### Poor Combined Performance
|
||||
**Analysis Steps**:
|
||||
1. Review individual strategy performance
|
||||
2. Check strategy correlation and overlap
|
||||
3. Adjust weights and combination rules
|
||||
4. Consider market regime compatibility
|
||||
|
||||
### Import Errors
|
||||
```
|
||||
ImportError: cannot import name 'StrategyManager'
|
||||
```
|
||||
**Solution**: Use correct import path: `from cycles.strategies import StrategyManager`
|
||||
|
||||
## 📞 Support
|
||||
|
||||
For issues, feature requests, or contributions:
|
||||
1. Check existing documentation and examples
|
||||
2. Review troubleshooting section
|
||||
3. Examine configuration files for proper syntax
|
||||
4. Ensure all dependencies are installed
|
||||
|
||||
---
|
||||
|
||||
**Version**: 1.0.0
|
||||
**Last Updated**: January 2025
|
||||
**TCP Cycles Project**
|
||||
Reference in New Issue
Block a user