Risk Management

Position sizing, stop losses, portfolio construction, and downside protection strategies are the foundation of any sustainable quantitative trading operation.

Key Points

• The Kelly criterion provides the mathematically optimal bet size as a fraction of capital, but most professionals use fractional Kelly (0.25x–0.5x) to reduce variance and avoid ruin risk
• Maximum drawdown targets should be set at the strategy level, not the portfolio level, and reversion-to-mean should be measured in weeks, not days
• Stop losses should be volatility-adjusted, not fixed-percentage; using ATR multiples keeps the strategy adaptive to changing market regimes
• Portfolio construction dominates individual strategy selection: combining uncorrelated return streams is the single highest-leverage activity
• Risk per trade is typically 0.5%–2% of capital; more than 2% on any single position is a sign of unmodeled edge
• The Sharpe ratio’s failure modes (non-normal returns, serial correlation) are addressable with the Sortino ratio and deflated Sharpe adjustments

Strategies

Volatility-Adjusted Position Sizing

Scale position size inversely with recent realized volatility so each trade risks a constant dollar amount. Use 10–20 day realized vol divided into a target risk per trade to get the position.

Drawdown-Aware Capital Allocation

Reduce total exposure after consecutive losing days, scale back up only after realized recovery. Cap exposure at 1.5x normal after a 5% drawdown event.

Hedged Carry

Pair directional carry trades with offsetting hedges in correlated instruments to neutralize beta while harvesting the carry differential.

Metrics & Formulas

• Kelly fraction: f* = (p × b − q) / b where p = win rate, b = win/loss ratio
• Sharpe ratio: (mean return − risk-free) / std(return); assume 252 trading days annualization
• Sortino ratio: like Sharpe but uses downside deviation only
• Maximum drawdown: peak-to-trough decline over a window
• Calmar ratio: CAGR / |max drawdown|

Tools & Resources

• Python pandas / numpy — Backtesting and position sizing math
• vectorbt — Vectorized backtesting with portfolio-level risk metrics
• riskfolio-lib — Portfolio optimization with HRP, mean-variance, and risk parity
• TA-Lib — ATR and other technical indicators for volatility-based stops