What Is the Sharpe Ratio?

What it means

The Sharpe ratio asks how much return a strategy produced for each unit of volatility it took on:

Sharpe = (average return − risk-free return) / standard deviation of returns

It is usually annualized by multiplying by the square root of the number of return samples per year. A higher Sharpe means smoother return per unit of risk, as measured by volatility.

Pitfalls

• Annualization choice matters. Daily, hourly or 5-minute returns need different multipliers, and markets don't trade around the clock.
• Volatility isn't all of risk. Rare large gaps, illiquidity and fat tails can be under-represented. Check drawdown too.
• Short samples are noisy. A high Sharpe over a few months of data says little.
• Optimized Sharpe is biased upward. The best of many tested settings usually looks better than it will later.

How TradingLead calculates it

TradingLead's backtesting analytics treat Sharpe as opt-in:

1. Annualization starts disabled with empty fields. You must enter samples per year, the expected interval between samples and an annual risk-free rate (default 0).
2. Returns are interval returns of candle-close equity. The initial-capital pseudo-interval is excluded, and at least two real intervals are needed.
3. Every interval must match the expected spacing. Irregular or session-gapped samples make Sharpe unavailable; they are not resampled or zero-filled. An optional user-assumed session schedule supports 1-240 minute session-aligned bars.
4. The formula uses the sample standard deviation (n − 1) and converts the annual risk-free rate to a per-sample rate:

rf_sample = (1 + riskFreeRateAnnual)^(1 / samplesPerYear) − 1
sharpe = (mean(r) − rf_sample) / stdev(r) × √samplesPerYear

If returns have zero variance, Sharpe is unavailable. Annualized volatility and the Sortino ratio use the same assumptions. Absolute return, P&L and drawdown are always available without them.

Reading it well

Compare Sharpe only between runs with the same sampling assumptions, look at it alongside drawdown and profit factor, and treat a backtest Sharpe as a description of the past, not a forecast.