volatility-based position sizing
When trading stocks systematically, using volatility-based position sizing offers a more robust, adaptive risk control framework than the traditional fixed stop-loss method. It aligns with quantitative rigor and improves portfolio stability.
Volatility-Based Position Sizing Explained
Volatility-based position sizing adjusts how much capital is allocated to a trade depending on the asset's recent volatility. The more volatile the asset, the smaller the position size. This ensures each trade carries a similar level of risk, regardless of price behavior.
Position Size = Risk Per Trade / (ATR × Multiplier)
- ATR: Average True Range (proxy for recent volatility)
- Multiplier: Often 1 to 2, adjusts sensitivity
- Risk Per Trade: Typically a fixed % of account equity (e.g., 1%)
Problems with Standard Stop-Losses
Standard stop-losses exit a trade at a fixed price or % level. While simple, they introduce several inefficiencies:
- Non-Adaptive: Doesn’t account for asset-specific or time-varying volatility
- Noise Sensitivity: Prone to premature exits due to random market noise
- Binary Outcomes: Results in cliff-edge risk profiles—either fully in or fully out
- Volatility Clustering: Misses changes in regime or market structure
Advantages of Volatility-Based Sizing
- Risk Normalization: All trades contribute evenly to portfolio risk
- Smoother Drawdowns: Avoids overexposure during volatile periods
- Consistency: More systematic, less emotionally driven decision-making
- Statistical Alignment: Integrates well with volatility models (e.g., GARCH)
Critique of Stop-Losses in Quantitative Literature
Authors like Robert Carver and Andreas Clenow argue that:
- Stop-losses may increase portfolio volatility
- They introduce distortion in backtests by masking true risk
- Exit signals should be based on predictive logic, not arbitrary price thresholds
Practical Implementation Without Stop-Losses
- Use volatility targeting to scale positions dynamically
- Exit based on trend reversals or signal decay
- Apply portfolio-level risk limits and drawdown controls rather than per-trade stops