Consider a hypothetical three-year record with annual returns +18%, −4% and +23%, no external cash flows and a stated maximum equity drawdown of 20%. The compounded equity factor is 1.393344, giving CAGR about 11.69% and Calmar about 0.585 under the convention used here. The drawdown is a separate path assumption: annual closing values alone cannot establish it.

What is the Calmar ratio

Under the convention used here, Calmar is compounded annual return divided by the positive magnitude of maximum drawdown from the same window. PerformanceAnalytics documents a traditional three-year window but allows other input lengths. State both the numerator definition and window; the name alone does not settle every reporting convention.

The denominator is the largest observed peak-to-trough decline, rather than the standard deviation of all returns. The ratio summarises annualised growth relative to that historical decline. It does not express a guaranteed cash return for each unit of future loss.

Three examples show how that works:

  • A system earning 12 percent a year with a maximum drawdown of 15 percent produces a Calmar ratio of 0.8.
  • For hypothetical CAGR 20% and MDD 25%, Calmar is 0.8. Matching the first example’s ratio does not make the two equity paths equivalent.
  • For hypothetical CAGR 9% and MDD 5%, Calmar is 1.8. This smaller observed drawdown is not proof of lower future risk or lower return volatility.

Formula and calculation method

The formula for the Calmar ratio is expressed as follows:

Calmar ratio = Compound annual return (CAGR) / Maximum drawdown

In this calculation, we take two components into account:

  • Compound annual growth rate (CAGR) is the annualised growth of the historical capital ratio under the stated cash-flow treatment. Without external cash flows, take (ending equity / starting equity)1 / years − 1. An exactly 39% three-year gain gives about 11.60% CAGR. The separate +18%, −4%, +23% example gives 39.3344% total growth and about 11.69% CAGR.
  • Maximum drawdown (MDD) is the largest percentage decline from a preceding running equity peak to a subsequent observation in the window, including an episode not yet recovered. State sampling frequency and treatment of external cash flows.

For comparison with a MAR-labelled report, inspect its window and numerator rather than assuming the label fixes them. A rolling-window calculation and an inception-to-date calculation can retain different drawdowns. In an inception-based report, annualised return and later drawdowns can still change even though an old episode remains in the data.

Two-year growth is not an annual return

Before dividing by drawdown, check that the report uses an annual compounded numerator. Total growth over two years answers a different question.

Scroll sideways to see the full diagram

Two-year growth is not an annual returnHypothetical equity path:100 units initially,90 after one year,144 after two. With linear paths between these points, maximum drawdown is 10%, from 100 to 90. Total growth is 44%; CAGR is the square root of 1.44 minus1, or20%. The article's Calmar convention gives 2, not4.4.Equity: 100 → 90 → 144Exactly 2 years; no cash flows80100120140100901440 years1 year2 yearsTotal: +44%; annual compounded: +20%MDD on this path: 10%Calmar: 20% ÷ 10% = 244% ÷ 10% = 4.4 uses another numerator.

The hypothetical model lasts exactly two years, with no external cash flows or additional fees. Equity uses one abstract unit. The entire path between the three points is assumed linear; MDD10% applies only to that disclosed path, not unknown intraday data. Morningstar defines this convention as compounded annual return divided by maximum drawdown.

This Calmar convention uses CAGR 20%, not the two-year44%.

Calculation: CAGR = (144 ÷ 100)1/2 − 1 = 20%. The shown MDD is(100 − 90) ÷ 100 = 10%, giving Calmar 2. Dividing44% by 10% gives 4.4 but uses total two-year return. The higher number does not improve the equity path; it changes the numerator's definition. Also check the observations used to measure drawdown.

Worked hypothetical three-year record

Return to the hypothetical record. The stated drawdown is an additional path assumption; the three annual returns alone cannot establish the deepest equity decline within those years.

Three hypothetical years without external cash flows
First-year result+18%
Second-year result−4%; no market-regime conclusion is implied
Third-year result+23%
Total growth in equity1.18 × 0.96 × 1.23 = 1.393344, that is +39.3344%
Compound annual return (CAGR)1.393344^(1/3) − 1 ≈ 11.69%
Maximum drawdown20% (peak to trough, in the second year)
Calmar ratio11.6913 / 20 ≈ 0.585
InterpretationA ratio under this model; suitability and strategy type do not follow from the ratio alone

The arithmetic mean of +18%, −4% and +23% is 12.3333%, giving about 0.617 when divided by 20%. CAGR is about 11.6913%, giving about 0.585. The geometric mean is no greater than the arithmetic mean for these positive capital factors, but neither ratio guarantees future return or drawdown.

Calmar compared with the Sharpe and Sortino ratios

The three ratios complement one another, and what separates them is how each one defines risk:

Performance ratios comparison
Sharpe ratioCompares return to overall volatility (measured by standard deviation)
Sortino ratioCompares mean return above a chosen target with downside deviation below that target
Calmar ratioCompares return to the largest historical equity drop (max drawdown)
Risk metric Sharpe/SortinoReturn dispersion: standard deviation for Sharpe; target-based downside deviation for Sortino
Risk metric CalmarLargest observed peak-to-trough percentage decline
Recommended practiceReporting all three metrics together

Benchmarks and performance assessment

The following ranges are illustrative reporting labels, not validated industry grades or universal targets. Compare periods, cash flows, costs and drawdown observations before judging a strategy.

Calmar ratio classification
Below 0.5Annual compounded return is less than half the stated maximum drawdown
Between 0.5 and 1.0Annual compounded return is between half and all of the stated maximum drawdown
Between 1.0 and 2.0Annual compounded return is one to two times the stated maximum drawdown
Between 2.0 and 5.0Annual compounded return is two to five times the stated maximum drawdown
Above 5.0Annual compounded return exceeds five times the stated maximum drawdown

A market or fund comparison needs named data, the exact observation window, total-return treatment, costs and drawdown sampling. This article supplies no verified index or managed-futures dataset for the proposed benchmark bands. A short favourable window can omit difficult paths; disclose the window rather than presenting an unverified industry norm.

Sample size and metric reliability

A three-year window is a documented convention, not a reliability guarantee. MDD depends on the path and observations within the selected window. Disclose what the history covers and omits rather than imposing an unsupported hierarchy of credible durations.

  • One-year period: report the observations and limitations. A short history may omit severe events; it is not automatically meaningless.
  • Three-year period: a traditional window, without assurance that different market phases or rare shocks are represented.
  • Five-year period: a longer observed record, not an absolute reliability minimum or guarantee of a full market cycle.
  • Ten-year period: additional history, not a complete picture of every future shock.

Disclose duration, sampling and cost treatment alongside the result. For example, a hypothetical Calmar of 1.2 over two years is a two-year calculation, not a claim that the same relationship will persist. A longer sample can add observed paths but cannot establish complete future coverage.

Best practices in performance reporting

A report can add complementary measures to show features Calmar omits. Use explicit definitions and a consistent data basis; reporting more ratios does not automatically make the underlying record reliable.

  1. Compound annual return — shows the overall profitability of the system.
  2. Maximum drawdown — determines the worst-case historical risk scenario.
  3. Sharpe ratio — evaluates return adjusted for overall volatility.
  4. Sortino ratio — analyzes return in relation to downside volatility.
  5. Calmar ratio — presents the relationship between return and maximum drawdown.
  6. Win rate and profit factor — summarise outcomes over the chosen sample; they do not diagnose every entry or guarantee future edge.
  7. Track record duration — provides context for all other figures.

Each of these numbers is easy to flatter on its own — pick the right period, or leave out the worst quarter. Put together, they keep each other honest: a high Calmar ratio next to a dismal profit factor shows up immediately, as does a fine win rate bought with a deep drawdown. A retail trader who wants to take their account seriously can keep this set in a single spreadsheet and update it once a month.

What to do tomorrow

  1. Collect equity observations as well as trades. A closed-trade export alone may miss intraday or floating drawdowns. Use the chosen sampling and cash-flow treatment to calculate CAGR and MDD from the same window.
  2. Incorporate Sharpe and Sortino calculations into your journal. Set up the mathematical formulas in your performance spreadsheet to automatically compute all three metrics after each trading week, providing you with a clearer perspective on your risk.
  3. Separate a review rule from the historical ratio. If a drawdown threshold triggers a review, record that threshold and the action explicitly. No universal Calmar 0.5 cut-off or 15% drawdown is established here, and a review rule cannot cap execution losses.

Related reading: Sharpe ratio — return adjusted for overall volatility; Sortino ratio — return adjusted for downside volatility; maximum drawdown — a key input parameter for the Calmar ratio formula.