OUTLAW-AI-Trading / code /advanced_trading_demo.py
BUBOUTLW's picture JonusNattapong's picture
Duplicate from JonusNattapong/AI-XAUUSD-Trading
43f8a47
Raw History Blame Contribute Delete
18.2 kB
#!/usr/bin/env python3
"""
Advanced Trading System Demo
Showcases trailing stops, profit taking, and dynamic exit strategies
"""
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from datetime import datetime, timedelta
import random
from market_regime_detector import MarketRegimeDetector
class AdvancedTradingSystem:
"""
Advanced trading system with trailing stops, profit taking, and dynamic exits
"""
def __init__(self, capital=1000, leverage=50):
self.capital = capital
self.leverage = leverage
# Trailing stop parameters - TIGHTENED for better profit capture
self.trailing_stop_pct = 0.025 # Reduced from 5% to 2.5% trailing stop
self.trailing_stop_distance = 0
# Profit taking parameters - MULTIPLE SCALED TARGETS
self.profit_targets = [0.01, 0.02, 0.05, 0.10] # 1%, 2%, 5%, 10% profit targets
self.take_profit_pct = 0.10 # 10% take profit
self.partial_take_profit_pct = 0.02 # Take partial profits at 2% (reduced from 5%)
# Breakeven stop parameters
self.breakeven_trigger_pct = 0.015 # Move to breakeven after 1.5% profit
self.breakeven_activated = False
# Initialize market regime detector
self.regime_detector = MarketRegimeDetector()
# Risk management
self.max_holding_time = 24 # Max hours
self.min_confidence = 0.3
# Performance tracking
self.trades = []
self.portfolio_values = []
# Current position
self.position = 0
self.entry_price = 0
self.entry_time = None
self.highest_price_since_entry = 0
def calculate_confidence(self, signal_strength, market_condition):
"""Calculate trading confidence"""
base_confidence = min(abs(signal_strength), 1.0)
market_multiplier = market_condition # 0.5-1.5
confidence = base_confidence * market_multiplier
return max(self.min_confidence, min(1.0, confidence))
def execute_trade(self, signal, confidence, price, timestamp):
"""Execute trade with advanced exit strategies"""
# Update regime parameters dynamically
self._update_regime_parameters(price, timestamp)
pnl = 0
exit_reason = None
# Check for exits first
if self.position != 0:
exit_reason = self._check_exits(price, timestamp, signal)
if exit_reason:
# Calculate P&L
if self.position > 0: # Long
pnl = (price - self.entry_price) * self.position
else: # Short
pnl = (self.entry_price - price) * abs(self.position)
self.capital += pnl
self._record_trade(timestamp, price, pnl, exit_reason, confidence)
# Reset position
self.position = 0
self.entry_price = 0
self.entry_time = None
self.trailing_stop_distance = 0
self.breakeven_activated = False # Reset breakeven flag
self.highest_price_since_entry = 0
# Enter new positions
elif abs(signal) > 0.1 and confidence > 0.2: # Lower thresholds
position_size = self._calculate_position_size(signal, confidence, price)
if signal > 0: # Buy
self.position = position_size
self.entry_price = price
self.entry_time = timestamp
self.highest_price_since_entry = price
self.trailing_stop_distance = price * (1 - self.trailing_stop_pct)
self.breakeven_activated = False # Reset breakeven flag
self._record_trade(timestamp, price, 0, 'entry', confidence, 'BUY')
elif signal < 0: # Sell
self.position = -position_size
self.entry_price = price
self.entry_time = timestamp
self.highest_price_since_entry = price
self.trailing_stop_distance = price * (1 + self.trailing_stop_pct)
self.breakeven_activated = False # Reset breakeven flag
self._record_trade(timestamp, price, 0, 'entry', confidence, 'SELL')
# Update trailing stops
if self.position != 0:
self._update_trailing_stops(price)
self.portfolio_values.append(self.capital)
return pnl, exit_reason
def _check_exits(self, price, timestamp, current_signal):
"""Check various exit conditions with improved profit-taking"""
# Calculate current profit percentage
if self.position > 0:
profit_pct = (price - self.entry_price) / self.entry_price
else:
profit_pct = (self.entry_price - price) / self.entry_price
# Breakeven stop activation
if not self.breakeven_activated and profit_pct >= self.breakeven_trigger_pct:
self.breakeven_activated = True
# Move trailing stop to breakeven + small buffer
buffer_pct = 0.005 # 0.5% buffer above breakeven
if self.position > 0:
self.trailing_stop_distance = self.entry_price * (1 + buffer_pct)
else:
self.trailing_stop_distance = self.entry_price * (1 - buffer_pct)
# Scaled profit taking - take partial profits at multiple levels
for target_pct in sorted(self.profit_targets, reverse=True):
if profit_pct >= target_pct:
# Calculate how much profit to take at this level
if target_pct <= 0.02: # Small profits (1-2%) - take 25% of position
profit_portion = 0.25
exit_reason = f'profit_{int(target_pct*100)}pct_partial'
elif target_pct <= 0.05: # Medium profits (5%) - take 50% of position
profit_portion = 0.50
exit_reason = f'profit_{int(target_pct*100)}pct_partial'
else: # Large profits (10%) - take full position
profit_portion = 1.0
exit_reason = 'take_profit'
if profit_portion < 1.0:
# Partial exit
self.position *= (1 - profit_portion)
return exit_reason
else:
# Full exit
return exit_reason
# Trailing stop check (only if breakeven not activated or profit is positive)
if self.position > 0 and price <= self.trailing_stop_distance:
return 'trailing_stop'
elif self.position < 0 and price >= self.trailing_stop_distance:
return 'trailing_stop'
# Maximum holding time - more aggressive for losing positions
if self.entry_time and (timestamp - self.entry_time).total_seconds() > self.max_holding_time * 3600:
return 'max_time'
# Early exit for significant losses (stop loss)
if profit_pct <= -0.03: # 3% stop loss
return 'stop_loss'
# Opposite signal (exit on strong reversal)
if (self.position > 0 and current_signal < -0.3) or (self.position < 0 and current_signal > 0.3):
return 'opposite_signal'
return None
def _update_regime_parameters(self, current_price, timestamp):
"""Update trading parameters based on current market regime"""
# Create a simple DataFrame with recent price history for regime detection
# In a real system, this would use more comprehensive market data
# For demo purposes, create synthetic OHLC data around current price
recent_prices = []
base_price = current_price
# Generate last 100 periods of synthetic data
for i in range(100):
# Add some realistic price movement
change = np.random.normal(0, 0.005) # 0.5% volatility
price = base_price * (1 + change)
recent_prices.append({
'Close': price,
'High': price * (1 + abs(np.random.normal(0, 0.002))),
'Low': price * (1 - abs(np.random.normal(0, 0.002))),
'Volume': np.random.randint(1000, 10000)
})
base_price = price
# Create DataFrame for regime detection
price_df = pd.DataFrame(recent_prices)
try:
# Detect current regime
current_regime, regime_params = self.regime_detector.detect_regime(price_df, -1)
# Update trading parameters based on regime
self.profit_targets = regime_params['profit_targets']
self.trailing_stop_pct = regime_params['trailing_stop_pct']
self.take_profit_pct = regime_params['profit_targets'][-1]
self.partial_take_profit_pct = regime_params['profit_targets'][1]
self.breakeven_trigger_pct = regime_params['breakeven_trigger']
self.max_holding_time = regime_params['max_holding_time']
# Store current regime
self.current_regime = current_regime
except Exception as e:
# If regime detection fails, use default parameters
self._set_default_parameters()
def _set_default_parameters(self):
"""Set default trading parameters"""
self.profit_targets = [0.01, 0.02, 0.05, 0.10]
self.trailing_stop_pct = 0.025
self.take_profit_pct = 0.10
self.partial_take_profit_pct = 0.02
self.breakeven_trigger_pct = 0.015
self.max_holding_time = 24
def _update_trailing_stops(self, price):
"""Update trailing stop levels"""
if self.position > 0: # Long position
if price > self.highest_price_since_entry:
self.highest_price_since_entry = price
self.trailing_stop_distance = price * (1 - self.trailing_stop_pct)
else: # Short position
if price < self.highest_price_since_entry:
self.highest_price_since_entry = price
self.trailing_stop_distance = price * (1 + self.trailing_stop_pct)
def _calculate_position_size(self, signal, confidence, price):
"""Calculate position size based on confidence"""
base_size = self.capital * self.leverage * min(abs(signal), 1.0) / price
confidence_multiplier = confidence ** 0.5 # Square root scaling
return base_size * confidence_multiplier
def _record_trade(self, timestamp, price, pnl, reason, confidence, action_type='EXIT'):
"""Record trade details"""
trade = {
'timestamp': timestamp,
'price': price,
'pnl': pnl,
'reason': reason,
'confidence': confidence,
'action': action_type,
'position_size': abs(self.position) if self.position != 0 else 0
}
self.trades.append(trade)
def generate_realistic_price_data(n_points=1000):
"""Generate realistic price data with trends and volatility"""
base_price = 2000
prices = [base_price]
timestamps = [datetime.now() + timedelta(hours=i) for i in range(n_points)]
# Generate price movements
for i in range(1, n_points):
# Trend component
trend = 0.001 * np.sin(i * 0.01) # Slow trend
# Volatility component
volatility = np.random.normal(0, 0.02)
# Momentum component
momentum = 0.1 * (prices[-1] - prices[-2]) / prices[-2] if len(prices) > 1 else 0
# Combine components
price_change = trend + volatility + momentum
new_price = prices[-1] * (1 + price_change)
# Ensure reasonable bounds
new_price = max(1800, min(2200, new_price))
prices.append(new_price)
return pd.DataFrame({
'timestamp': timestamps,
'price': prices
})
def generate_trading_signals(price_data):
"""Generate realistic trading signals"""
signals = []
confidences = []
for i, row in price_data.iterrows():
price = row['price']
# Simple trend signal
if i > 20:
short_ma = price_data['price'].iloc[i-10:i].mean()
long_ma = price_data['price'].iloc[i-20:i].mean()
trend_signal = 2 * (short_ma - long_ma) / long_ma # Amplified signal
else:
trend_signal = 0
# Volatility-based confidence
if i > 10:
returns = price_data['price'].iloc[i-10:i].pct_change()
volatility = returns.std()
market_condition = max(0.5, min(1.5, 1 / (1 + volatility * 10)))
else:
market_condition = 1.0
# Add some noise to signals
noise = np.random.normal(0, 0.2) # Increased noise
signal = trend_signal + noise
confidence = min(1.0, abs(signal) * market_condition)
signals.append(signal)
confidences.append(confidence)
price_data['signal'] = signals
price_data['confidence'] = confidences
return price_data
def run_advanced_trading_demo():
"""Run comprehensive demo of advanced trading features"""
print("🚀 Advanced Trading System Demo")
print("=" * 50)
print("Features:")
print("• Confidence-based position sizing")
print("• Trailing stops (5%)")
print("• Profit taking (5% partial, 10% full)")
print("• Maximum holding time (24 hours)")
print("• Opposite signal exits")
print()
# Generate data
price_data = generate_realistic_price_data(500)
trading_data = generate_trading_signals(price_data)
# Initialize trading system
trader = AdvancedTradingSystem(capital=1000, leverage=50)
# Execute trades
print("Executing trades...")
for i, row in trading_data.iterrows():
pnl, exit_reason = trader.execute_trade(
row['signal'],
row['confidence'],
row['price'],
row['timestamp']
)
if i % 100 == 0:
print(f"Processed {i}/{len(trading_data)} data points...")
print("\n✅ Trading simulation completed!")
print(f"Final Portfolio Value: ${trader.capital:.2f}")
# Analyze results
analyze_results(trader)
# Plot results
plot_advanced_results(trader, trading_data)
def analyze_results(trader):
"""Analyze trading performance"""
if not trader.trades:
print("No trades executed")
return
df_trades = pd.DataFrame(trader.trades)
# Separate entries and exits
entries = df_trades[df_trades['action'].isin(['BUY', 'SELL'])]
exits = df_trades[df_trades['action'] == 'EXIT']
print("\n📊 Performance Analysis:")
print(f"Total Trades: {len(entries)}")
print(f"Total Exits: {len(exits)}")
if len(exits) > 0:
profitable_trades = len(exits[exits['pnl'] > 0])
win_rate = profitable_trades / len(exits) * 100
avg_win = exits[exits['pnl'] > 0]['pnl'].mean() if profitable_trades > 0 else 0
avg_loss = exits[exits['pnl'] < 0]['pnl'].mean() if len(exits[exits['pnl'] < 0]) > 0 else 0
print(f"Win Rate: {win_rate:.1f}%")
print(f"Average Win: ${avg_win:.2f}")
print(f"Average Loss: ${avg_loss:.2f}")
# Exit reason analysis
exit_reasons = exits['reason'].value_counts()
print("\nExit Reasons:")
for reason, count in exit_reasons.items():
pct = count / len(exits) * 100
print(f" {reason}: {count} ({pct:.1f}%)")
def plot_advanced_results(trader, price_data):
"""Plot comprehensive results"""
fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, figsize=(15, 10))
# Price chart with trades
ax1.plot(price_data['timestamp'], price_data['price'], alpha=0.7)
# Plot trades
if trader.trades:
df_trades = pd.DataFrame(trader.trades)
entries = df_trades[df_trades['action'].isin(['BUY', 'SELL'])]
buy_signals = entries[entries['action'] == 'BUY']
sell_signals = entries[entries['action'] == 'SELL']
ax1.scatter(buy_signals['timestamp'], buy_signals['price'],
color='green', marker='^', s=100, label='Buy')
ax1.scatter(sell_signals['timestamp'], sell_signals['price'],
color='red', marker='v', s=100, label='Sell')
ax1.set_title('Price Chart with Trade Signals')
ax1.set_ylabel('Price')
ax1.legend()
ax1.grid(True)
# Portfolio value over time
ax2.plot(trader.portfolio_values)
ax2.set_title('Portfolio Value Over Time')
ax2.set_ylabel('Portfolio Value ($)')
ax2.grid(True)
# P&L distribution by exit reason
if trader.trades:
df_trades = pd.DataFrame(trader.trades)
exits = df_trades[df_trades['action'] != 'entry']
if len(exits) > 0:
exit_reasons = exits['reason'].unique()
pnl_by_reason = [exits[exits['reason'] == reason]['pnl'] for reason in exit_reasons]
ax3.boxplot(pnl_by_reason, tick_labels=exit_reasons)
ax3.set_title('P&L Distribution by Exit Reason')
ax3.set_ylabel('P&L ($)')
ax3.grid(True)
# Confidence vs P&L
if trader.trades:
df_trades = pd.DataFrame(trader.trades)
exits = df_trades[df_trades['action'] != 'entry']
if len(exits) > 0:
colors = ['green' if x > 0 else 'red' for x in exits['pnl']]
ax4.scatter(exits['confidence'], exits['pnl'], c=colors, alpha=0.6)
ax4.set_title('Confidence vs P&L')
ax4.set_xlabel('Confidence')
ax4.set_ylabel('P&L ($)')
ax4.grid(True)
plt.tight_layout()
plt.savefig('advanced_trading_demo.png', dpi=300, bbox_inches='tight')
plt.show()
if __name__ == "__main__":
run_advanced_trading_demo()