Enhanced-Kalman-Filter-Statistical-Arbitrage-Engine

📈 Kalman Pairs Trading Engine

A statistical arbitrage framework using Kalman Filter to dynamically estimate hedge ratios between cointegrated asset pairs. Enhanced with:

• ✅ Volatility Scaling
• ✅ Dynamic Z-score Thresholds
• ✅ Slippage Modeling
• ✅ Risk-based Position Sizing

---

🔢 Performance Highlights (KO vs. PEP)

Metric	Value
Sharpe Ratio	1.30
Max Drawdown	-7.72%
CAGR (Simulated)	18.5%
Trade Accuracy	63.2%

---

🧠 Methodology

[ z_t = \frac{(y_t - \beta_t x_t - \alpha_t)}{\sigma_t}, \quad \text{where } [\beta_t, \alpha_t] \sim \text{KalmanFilter} ]

Trade signals are generated based on dynamic thresholds:

• Enter Long: ( z_t < -2 \cdot \frac{\sigma_t}{\mu_t} )
• Exit: ( |z_t| < 0.5 \cdot \frac{\sigma_t}{\mu_t} )

---

📊 Strategy Outputs

• 🔄 Dynamic hedge ratio plots
• 💰 PnL and equity curve
• 🧪 Entry/exit signals marked on spread charts

---

🚀 How to Run

pip install -r requirements.txt
python run_strategy.py --csv data/example_pairs.csv