Sharpe Ratio
Return per unit of total risk (volatility); the standard risk-adjusted performance metric for backtesting.
Definition
The Sharpe ratio is (R − R_f) / σ, where R is the strategy’s mean return, R_f is the risk-free rate, and σ is the standard deviation of returns (volatility). It measures excess return per unit of total risk.
Annualized form: Sharpe_ann ≈ (mean excess return × √N) / σ_ann for N periods per year.
Why it matters for backtesting
- Comparability: Lets you compare strategies with different volatilities on a common scale.
- Risk-adjusted view: A high raw return with very high vol may have a lower Sharpe than a moderate return with low vol.
- Overfitting signal: Extremely high backtest Sharpes (e.g. > 2) often don’t persist; use with caution.
Limitations
- Assumes returns are roughly normal; underestimates tail risk.
- Sensitive to the risk-free rate and the period used.
- Single factor (volatility); does not distinguish upside vs downside vol.
Linked concepts
Sortino ratio (downside vol), Information ratio (active return vs tracking error), max drawdown, Calmar ratio.