This repository features a fully implemented and optimized market-neutral pairs trading strategy between GOOG and MSFT. The model uses Johansen cointegration, dynamic volatility targeting, ATR-based stop-loss, and parameter optimization to generate trades based on relative mispricing.
- ✅ Cointegration Detected (Johansen Test): GOOG & MSFT
- ✅ Fully parameterized backtest framework with Grid, Random & Bayesian optimization
- ✅ Robust risk control: Max drawdown capped at –8.7% with volatility filters & stop-loss
- ✅ Out-of-sample generalization confirms real-world applicability (2023–2024)
- ✅ Execution-aware modeling: Includes slippage, commission, and realistic position sizing using dynamic leverage
| Metric | Value |
|---|---|
| Total Return | 44.80% |
| Annualized Return | 6.27% |
| Annualized Volatility | 9.21% |
| Sharpe Ratio | 0.68 |
| Sortino Ratio | 0.78 |
| Max Drawdown | –8.67% |
⚖️ Lower return than Buy‑and‑Hold, but delivers superior risk control and market‑neutral alpha.
| Metric | Value |
|---|---|
| Total Return | 9.33% |
| Annualized Return | 8.64% |
| Annualized Volatility | 10.75% |
| Sharpe Ratio | 0.80 |
| Sortino Ratio | 0.67 |
| Max Drawdown | –8.87% |
| Metric | Value |
|---|---|
| Total Return | 275.68% |
| Annualized Return | 24.32% |
| Annualized Volatility | 28.74% |
| Sharpe Ratio | 0.85 |
| Sortino Ratio | 1.15 |
| Max Drawdown | –40.54% |
🚨 Strong returns due to bull market, but with extreme volatility and –40% drawdown.
Not suitable for long-term or conservative asset management—no downside hedge, no market neutrality.
A complete write-up including methodology, charts, code explanation, and interpretations:
👉 📄 View Full Report (Google Drive)
This is a realistic, risk-managed, self-built strategy suitable for use in asset management or hedge fund environments. It showcases:
- Practical Python implementation without external trading APIs
- Full-cycle model development (data → strategy → backtest → evaluation)
- A strong emphasis on overfitting prevention, capital preservation, and real-world deployability
- The strategy can serve as a foundation for intraday ETF pairs, stat-arb equity models, or portfolio hedging modules in live environments.
💼 Designed and implemented from scratch by a third-year FinTech student with real asset management experience (HKD 320,000 AUM), emphasizing practical risk control and self-learned quant research.