The Calmar Ratio: Return Relative to Maximum Drawdown

Every risk-adjusted return figure has to choose what it means by risk. One family measures the spread of returns; another measures the worst decline the equity curve actually suffered. The Calmar ratio belongs to the second family. It divides a strategy's annualized return by its maximum drawdown, so its idea of risk is the deepest hole, not the average swing. This primer defines the ratio, walks its arithmetic on an illustrative example, and shows where a drawdown-based measure is honest and where it can mislead.

By Imperial Analytics

What the Calmar ratio is

The numerator is the strategy's annualized return, the growth rate the equity curve produced expressed on a yearly basis. The denominator is the maximum drawdown over the same window, the single largest drop from an equity high-water mark down to the lowest point that followed before a new high was made. Both are usually stated as percentages, so the ratio comes out as a plain number with no units, comparable across strategies of different sizes.

The name records where it came from. Terry Young, who ran a managed-accounts firm and published a newsletter under its initials, built the measure to rate trading programs, and "Calmar" is drawn from that firm and newsletter name rather than from any statistical term. Its purpose was practical from the start: give an investor one figure that pairs the return a program earned with the worst loss an investor would have had to sit through to earn it.

How the Calmar ratio is built

Take an illustrative track record. Suppose a strategy earned an annualized return of twenty-four percent over a window, and the deepest peak-to-trough decline in its equity over that same window was twelve percent. The Calmar ratio is twenty-four divided by twelve, which is 2.0. Read plainly, the strategy returned twice its worst decline per year: for every unit of drawdown an investor had to endure, two units of annual return came back.

The window is part of the definition, not a detail. Young's original measure looked back over a fixed span, the trailing thirty-six months, and recomputed each month, so the ratio always described the recent past rather than the entire track record. That choice keeps the figure responsive to how a strategy is behaving lately, and it means a Calmar ratio is only meaningful when the window it was measured over is stated alongside it. The same return and the same equity curve can produce different ratios depending on how far back the drawdown is measured.

Why maximum drawdown is the risk term

Maximum drawdown measures path, not spread. It walks the equity curve in order, tracks the running high-water mark, and records the largest percentage fall from a peak to the trough that came before the next peak. Because it reads the sequence rather than a summary statistic, it captures something a trader experiences in real time: the deepest the account went underwater, and the loss they would have had to hold through without abandoning the strategy. A fuller treatment lives in the primer on maximum drawdown.

That choice of risk term is what gives the Calmar ratio its intuitive grip. A drawdown is the loss that tests whether a trader keeps trading a strategy at all; it is the number that decides position size, because the account has to survive the worst stretch to collect the average. By dividing return by exactly that figure, the Calmar ratio speaks in the units a trader already worries about. A ratio of 2.0 says the annual return was twice the worst decline; a ratio below 1.0 says the strategy at some point gave back more than a year of its own return in a single slide.

What the Calmar ratio does not tell you

The denominator rests on one observation. Maximum drawdown is the worst single decline in the record, which means the entire ratio can turn on a single stretch of the equity curve. Add more history and the maximum drawdown can only stay the same or grow, because a longer sample has more chances to contain a deeper slide; so a Calmar ratio measured over a short calm window flatters a strategy that has simply not met its bad stretch yet. This is the same sample-dependence that makes any single worst-case figure an unstable lower bound rather than a fixed property.

The ratio is also silent about the shape and rhythm of the risk. It uses the depth of the worst drawdown and nothing else, so it cannot tell a trader whether that decline arrived in one sharp week or ground on for months, nor how frequently smaller drawdowns occurred. Two strategies with the same annualized return and the same maximum drawdown score identically, even if one recovered in days and the other stayed underwater for a year. Depth is only one dimension of a drawdown; duration and frequency are invisible to the Calmar ratio, and they matter to whether a strategy is tradable.

How the Calmar ratio differs from its near neighbors

The nearest neighbor is the Sharpe ratio, which also divides a return by a measure of risk. The difference is the denominator. The Sharpe ratio uses the standard deviation of returns, an average of how far returns spread in both directions, so it penalizes a strategy's biggest up months as much as its down months. The Calmar ratio uses maximum drawdown, a one-sided measure of the worst realized loss, so it ignores upside variability and focuses entirely on the deepest decline. A strategy with a few large winning months can look worse on Sharpe and better on Calmar, because the two disagree about whether a big gain is a form of risk.

The other neighbors sit on the return side. Profit factor and expectancy describe the size of an edge: gross wins over gross losses, or the average result per trade. They add up or average outcomes and discard the order in which they arrived, so they say nothing about the path or the worst decline along it. Maximum drawdown on its own reports the path risk but not the return that came with it. The Calmar ratio combines the two sides, return over worst decline, which is its whole reason to exist and also why it should be read next to the pieces it summarizes rather than in place of them.

How to read and log the Calmar ratio without leaning on one number

The first discipline is the window. Because the maximum drawdown grows with the length of the record, a Calmar ratio has no meaning without the span it was measured over. Note whether it covers the trailing three years, the full track record, or a shorter stretch, and never compare two ratios measured over different windows as though they were the same quantity. Writing the window down beside the number is the equivalent of writing an invalidation level down with a trade: it is the context that makes the figure legible later.

The second discipline is the sample. A Calmar ratio computed over a handful of months is noise, because both the annualized return and the worst drawdown are unstable until enough time and enough trades accumulate. The same twenty-trade-per-setup floor that governs any pattern claim applies to the history behind a Calmar ratio, and the sample size a win rate needs is the same kind of check: a flattering ratio on a short, calm sample is the least trustworthy version of the number.

The third discipline is pairing. Read the Calmar ratio next to the Sharpe ratio and next to the raw maximum drawdown it was built from. If the two risk-adjusted figures disagree, the gap usually points at the shape of the returns, and looking at the actual drawdown resolves it. A Calmar ratio that drifts down over time, read against edge decay, can be an early sign a strategy is weakening; but a Calmar ratio read alone, with no window, no sample size, and no drawdown beside it, is a single number standing in for a risk it only partly describes.

Frequently asked questions

• q: What is a good Calmar ratio? a: There is no universal threshold, because the ratio depends on the window it was measured over and the sample behind it. As a rough reading, a ratio above 1.0 means the annualized return exceeded the worst drawdown, and higher is generally read as better return for the worst decline endured, but the figure is only trustworthy when the window and sample size are stated alongside it.
• q: What window does the Calmar ratio use? a: The original definition used a trailing thirty-six-month window, recomputed monthly, so the ratio describes recent return against the recent worst drawdown rather than the entire track record. Because maximum drawdown grows with the length of the record, the window is part of the definition and has to be stated for the number to be comparable.
• q: How is the Calmar ratio different from the Sharpe ratio? a: Both divide a return by a measure of risk, but the denominators differ. The Sharpe ratio uses the standard deviation of returns, a two-sided measure of variability that penalizes large gains as well as losses. The Calmar ratio uses maximum drawdown, a one-sided measure of the single worst decline, so it ignores upside variability and focuses on the deepest realized loss.
• q: Why does the Calmar ratio get worse with more data? a: Its denominator is the maximum drawdown, the single worst decline in the record. A longer sample has more chances to contain a deeper decline, so the maximum drawdown can only stay the same or grow as history is added. A ratio measured over a short calm window can look strong simply because the strategy has not yet met its worst stretch.

The Calmar ratio earns its place by speaking in the units a trader feels: return set against the worst decline the strategy actually dug. That is its strength and its limit in the same breath. Because the whole figure can rest on one drawdown, and because that drawdown grows with the length of the record, the ratio is only as honest as the window and sample stated beside it. Read with those two facts in view, and paired with a variability measure and the raw drawdown, it is a clear summary of return for pain. Read alone, it is one observation wearing the confidence of a law.

Sources

Footnotes

1. Terry W. Young, "Calmar Ratio: A Smoother Tool," Futures, Vol. 20, No. 1, 1991, p. 40. ↩