Sharpe Ratio

The Sharpe ratio measures the risk-adjusted return of a strategy. It tells you how much excess return you earn per unit of risk (volatility). A higher Sharpe ratio means better returns relative to the risk taken.

The Formula

Sharpe Ratio = (Strategy Return - Risk-Free Rate) / Standard Deviation of Returns

Where:

  • Strategy Return: Your annualized return
  • Risk-Free Rate: The return with zero risk (typically Treasury bill rate, ~4-5% recently)
  • Standard Deviation: The volatility of your returns (annualized)

In plain terms: if two strategies both return 20% annually but one has twice the volatility, the less volatile one has double the Sharpe ratio — it achieved the same return with less risk.

Quick Reference

Sharpe RatioInterpretation
< 0Losing money or underperforming risk-free rate
0.0 - 0.5Poor — barely compensated for risk
0.5 - 1.0Acceptable — typical for simple strategies
1.0 - 2.0Good — strong risk-adjusted performance
2.0 - 3.0Excellent — very few strategies sustain this
> 3.0Exceptional or suspicious — verify for errors

Most successful systematic funds run at Sharpe 1.0-2.0 net of fees. If your backtest shows Sharpe 4.0, you’re almost certainly overfit.

Why Sharpe Ratio Matters for Backtesting

The Sharpe ratio is the most widely used metric for evaluating backtest quality. Raw returns are misleading without context — Sharpe provides that context.

  • Apples-to-apples comparison: A 30% return with 40% volatility (Sharpe ~0.6) is worse risk-adjusted than a 15% return with 10% volatility (Sharpe ~1.0). The second strategy, leveraged 2x, would return ~30% with ~20% volatility — still better.
  • Scalability: High Sharpe = efficient risk-to-return conversion. A high-Sharpe strategy can be leveraged up; a low-Sharpe strategy can’t.
  • Overfitting detection: Sharpe above 3.0 for daily strategies is extremely rare in practice. Suspiciously high Sharpe is a red flag.
  • Better optimization target: Sharpe penalizes volatility, making overfit solutions less likely than optimizing raw return.

Concrete Examples

Comparing Two Strategies

Strategy A: 25% annual return, 30% volatility, risk-free rate 4%

Sharpe = (25% - 4%) / 30% = 0.70

Strategy B: 15% annual return, 10% volatility, risk-free rate 4%

Sharpe = (15% - 4%) / 10% = 1.10

Strategy B is superior risk-adjusted. Leverage Strategy B to 3x: ~45% return with 30% volatility — still a better risk profile than A.

The Sharpe Ratio Trap

A backtest on daily data from 2010-2020 (mostly bullish) shows Sharpe 2.5. The trader deploys capital. In 2022’s bear market, Sharpe drops to 0.3. The backtest Sharpe was inflated by testing on a favorable period.

Always test across multiple market regimes and report the time period.

Frequency and Sharpe

A strategy trading daily shows Sharpe 0.8. The same logic on monthly data: Sharpe 0.4. Higher-frequency strategies tend to show higher Sharpe because there are more opportunities and the law of large numbers works faster. Be careful comparing Sharpe across different trading frequencies.

Annualization

You calculate Sharpe on monthly returns and get 0.35. To annualize:

Annual Sharpe = Monthly Sharpe x sqrt(12) = 0.35 x 3.46 = 1.21

Always annualize before comparing. Note: sqrt(N) scaling assumes returns are independent. If returns are autocorrelated (common in trend-following), annualized Sharpe can be misleading.

Calculating Sharpe Correctly

  1. Use consistent return types: Either log returns or simple returns throughout. Don’t mix them.
  2. Annualize properly: Daily returns: multiply by sqrt(252). Monthly: sqrt(12). Weekly: sqrt(52).
  3. Choose the right risk-free rate: Use the prevailing short-term treasury rate. Setting it to 0 is common but state your assumption.
  4. Use sufficient data: Sharpe from 20 trades is statistically meaningless. Need at least 2-3 years or 100+ trades.
  5. Report confidence intervals: Sharpe of 1.0 from 3 years of data has standard error ~1/sqrt(3) = 0.58. True Sharpe could be anywhere from 0.4 to 1.6.

Sharpe Ratio Limitations

  • Assumes normal distribution: Treats upside and downside volatility equally. A strategy with occasional large gains gets penalized the same as one with large losses. Use the Sortino ratio (only penalizes downside) as a complement.
  • Ignores tail risk: Great Sharpe but occasional blowups (fat tails). Check kurtosis and skewness.
  • Time-period dependent: Varies significantly by test period. Report the period.
  • Gaming the Sharpe: Strategies selling options (small premiums, large tail risk) show artificially high Sharpe until a blowup occurs. Selling puts looks great — until it doesn’t.

Complementary Metrics

Don’t rely on Sharpe alone. Use alongside:

  • Sortino Ratio: Penalizes downside volatility only — better for asymmetric strategies
  • Calmar Ratio: Return / Max Drawdown — focuses on worst case
  • Profit Factor: Gross Profit / Gross Loss — intuitive measure of edge
  • Win Rate + Payoff Ratio: How often you win and how big wins are vs. losses
  • Maximum Drawdown: Absolute worst case regardless of volatility

Resources