Volatility-Weighted Multi-Strategy Portfolio (Python)

A systematic multi-strategy research project that combines three market-neutral long/short strategies and allocates capital using inverse-volatility and risk-parity-style weighting, including transaction cost modeling.

Strategies

1. Cross-Sectional Momentum (Long/Short)

   • Ranks assets by risk-adjusted trailing return (return / volatility)
   • Long top quantile, short bottom quantile (market-neutral)

2. Cross-Sectional Mean Reversion (Long/Short)

   • Short-horizon reversal: long recent losers, short recent winners
   • Quantile-based market-neutral construction

3. Breakout Channel (Long/Short)

   • Uses a rolling price channel to score strength/weakness
   • Quantile-based market-neutral construction

Portfolio Construction

The strategy return streams are combined using one of:

• inverse_vol: weights ∝ 1 / rolling volatility
• risk_parity: rolling covariance-based risk parity (iterative equal risk contribution)
• *_sharpe_filter: optionally down-weights strategies with negative rolling Sharpe

Strategy weights are applied with a 1-day lag to avoid look-ahead bias.

Costs

Transaction costs are modeled as:

• commission + slippage (bps)
• applied proportional to turnover at rebalances (strategy level and portfolio level)

Repo Structure

multi-strategy-portfolio/
├── requirements.txt
├── config.py
├── data.py
├── costs.py
├── metrics.py
├── portfolio.py
├── backtest.py
├── run.py
└── strategies
├── init.py
├── momentum.py
├── mean_reversion.py
└── breakout.py