Quantitative Portfolio Optimization & Rolling Backtest

Overview

This project implements a full quantitative portfolio research pipeline including:

• Data download
• Return calculation
• Covariance & correlation analysis
• Monte Carlo portfolio optimization
• Efficient Frontier visualization
• Rolling out-of-sample backtest
• Risk metrics (Sharpe, Sortino, VaR, CVaR)
• Benchmark comparison

Methodology

• Optimization Objective: Maximize Sharpe Ratio
• Rolling Window: 5-year training, 1-year test
• Rebalancing: Annual
• Monte Carlo Simulations: 2000 per rebalance

Key Results

• Rolling CAGR: 25%+
• Rolling Sharpe: ~1.0
• Sortino Ratio: > 1
• Outperformance vs SPY benchmark

Structure

src/ → Research code
data/raw/ → Raw downloaded data
data/processed/ → Processed returns
data/outputs/ → Charts & results

How To Run

pip install -r requirements.txt
python src/rolling_backtest.py
python src/risk_metrics.py

---

Author: Indrajith