Measuring a strategy honestly

A single headline return tells you almost nothing. These pages define every metric a professional uses to judge a backtest — return, risk-adjusted return, drawdown and edge measures — with the exact formula, how to read it, how professionals actually use it, and, crucially, the ways each number can mislead you. No metric is trusted alone; the craft is reading them together.

Performance Metrics: Backtest performance metrics fall into families: return (CAGR, absolute and annualized return), risk-adjusted return (Sharpe, Sortino, Calmar, information ratio), drawdown and pain (maximum drawdown, average drawdown, recovery factor, Ulcer index), and edge (profit factor, win rate, payoff ratio, expectancy). Each has a precise formula and a specific blind spot, so they are read together and always alongside the drawdown that shows how much pain earned the return.

CAGR (Compound Annual Growth Rate)

Return metric

CAGR (compound annual growth rate) is the single constant yearly rate that, compounded over the full test period, would take the starting equity to t…

Absolute Return

Return metric

Absolute return is the total percentage change in equity from the start to the end of a backtest, computed as (End − Start) ÷ Start, with no adjustme…

Annualized Return

Return metric

Annualized return is a period return scaled to a full-year rate by geometric compounding, computed for a per-period return r over p periods per year …

Sharpe Ratio

Risk-adjusted

The Sharpe ratio is a risk-adjusted performance measure equal to the portfolio's excess return over the risk-free rate divided by the standard deviat…

Sortino Ratio

Risk-adjusted

The Sortino ratio is a risk-adjusted measure equal to the portfolio's excess return over a target divided by the downside deviation, the standard dev…

Calmar Ratio

Risk-adjusted

The Calmar ratio is a risk-adjusted measure equal to the compound annual growth rate divided by the absolute value of the maximum drawdown, expressin…

Information Ratio

Risk-adjusted

The information ratio is a risk-adjusted measure equal to the portfolio's active return over a benchmark divided by the tracking error, the standard …

Maximum Drawdown

Risk metric

Maximum drawdown is the largest percentage decline from a historical peak to a subsequent trough in an equity curve, taken as the maximum over all po…

Average Drawdown

Risk metric

Average drawdown is the mean depth of the declines below the high-water mark across an equity curve, describing the typical pain a strategy inflicts …

Recovery Factor

Risk metric

The recovery factor is a risk metric equal to a strategy's net profit divided by the absolute value of its maximum drawdown, measuring how many times…

Profit Factor

Trade metric

The profit factor is a trade-level efficiency metric equal to the gross profit from all winning trades divided by the gross loss from all losing trad…

Win Rate

Trade metric

The win rate is the fraction of trades that close profitable, computed as the number of winning trades divided by the total number of trades, describ…

Payoff Ratio

Trade metric

The payoff ratio is a trade-level metric equal to the average winning trade divided by the average losing trade, describing how large the strategy's …

Expectancy

Trade metric

Expectancy is the average profit or loss a strategy produces per trade, computed as the win rate times the average win minus the loss rate times the …

Ulcer Index

Risk metric

The ulcer index is a drawdown-based risk measure equal to the square root of the mean of the squared percentage drawdowns at every point in time, cap…

Frequently asked questions

What is the most important backtest metric?
There is no single most important metric. Return figures like CAGR are meaningless without a risk measure like maximum drawdown, and risk-adjusted ratios like Sharpe or Sortino combine the two. Professionals read a return metric, a risk-adjusted ratio and a drawdown measure together, and weight the drawdown heavily because it decides survival.
What is a good Sharpe ratio for a backtest?
As a rough guide, a long-run Sharpe near or above 1 is often considered solid and above 2 excellent, but the figure depends heavily on the strategy, sample length and costs. Backtested Sharpe ratios almost always fall once realistic frictions and out-of-sample data are included, so treat a high in-sample Sharpe with suspicion.
Why isn't a high win rate enough?
Win rate ignores the size of wins and losses. A strategy that wins 70% of the time but loses far more on its losers than it makes on its winners can still have negative expectancy. Win rate must be read with the payoff ratio, or combined into expectancy, to mean anything.
Educational content only — not investment advice. See our Risk Disclosure.