The Calmar ratio is your annualised return divided by your maximum drawdown: how much return you earned for the worst peak-to-trough fall your account suffered. A Calmar of 3 means a year of returns at your recent pace is three times the worst decline. Where the Sharpe ratio and Sortino ratio measure return against day-to-day swings, Calmar measures it against the one loss you could not trade through. That is the number a drawdown limit at a prop firm is built on. This page works the Calmar ratio on the same ten trading days as those two posts, then on two traders with identical profit and very different worst drawdowns.
The Calmar ratio formula
Calmar ratio = annualised return ÷ maximum drawdown
- Annualised return: your average daily return × 252 trading days. This page and ExpectancyIQ both use this simple scaling, not compounding. Each day's return is that day's P&L divided by the balance going into it.
- Maximum drawdown: the largest fall from a running peak to a later low, as a percentage of that peak, measured on your account's compounded daily return curve. See max drawdown.
- Undefined with no drawdown. If the account has never fallen below a prior peak, the denominator is zero and the ratio does not exist. ExpectancyIQ shows a dash, not 0 and not infinity.
The ratio is usually credited to Terry Young, who named it after his newsletter, California Managed Accounts Reports. Most published definitions use a trailing 36-month window for both the return and the drawdown. I could not check Young's original article, so treat that history and the 36-month convention as secondhand. A related ratio, often called the MAR ratio, uses the whole history instead of 36 months, and some software differs on whether drawdown is measured monthly or daily. As with Sortino, confirm which variant a tool uses before comparing numbers.
A Calmar ratio worked example with 10 trading days
These are the same ten daily returns used in the Sharpe ratio example. The drawdown column is how far the compounded account sits below its highest point so far, at the close of each day.
| Day | Return | Drawdown |
|---|---|---|
| 1 | +0.8% | 0 |
| 2 | −0.4% | 0.40% |
| 3 | +1.2% | 0 |
| 4 | +0.3% | 0 |
| 5 | −0.9% | 0.90% |
| 6 | +0.6% | 0.31% |
| 7 | 0.0% | 0.31% |
| 8 | +1.1% | 0 |
| 9 | −0.5% | 0.50% |
| 10 | +0.7% | 0 |
- Average daily return: 0.29%
- Annualised: 0.29% × 252 = 73.08%
- Maximum drawdown: the deepest row above, on day 5, = 0.90%
- Calmar: 73.08 ÷ 0.90 = 81.2
An 81 is not a finding. Annualising ten days turns a 2.9% gain into 73%, and the denominator is one bad day. The mechanics are the point here; the sample-size section below says why the number cannot be trusted yet.
Calmar ratio on the same profit, different drawdowns
These are the three traders from the Sortino post. Each made +2.9% over ten days, an average of 0.29% a day, so every one annualises to the same 73.08%. Only the path differs.
- Trader A is the series above: small moves, worst fall 0.90%.
- Trader B made big swings both ways: +2.4%, −1.8%, +3.0%, −0.6%, −2.7%, +1.9%, −1.2%, +2.8%, −1.6%, +0.7%. Worst fall 3.28%.
- Trader C lost a little most days and won big twice: +3.2%, −0.3%, −0.2%, −0.5%, +2.6%, −0.4%, −0.3%, −0.2%, −0.5%, −0.5%. Worst fall 1.89%, from five small losses in a row.
| Trader | Max drawdown | Calmar |
|---|---|---|
| A | 0.90% | 81.2 |
| B | 3.28% | 22.3 |
| C | 1.89% | 38.7 |
Here Calmar sees something Sortino cannot. In the Sortino comparison A and C tie at 12.2, because their downside deviation is the same 0.36%. But C's losses arrive in a row, and a drawdown adds consecutive losses together. C ends up with twice A's worst fall, so its Calmar is less than half. Sortino measures how deep each bad day is. Calmar measures how deep the bad stretch is.
Calmar ratio and prop firm drawdown limits
A funded or evaluation account usually ends at a drawdown limit, so max drawdown is the figure that decides whether you keep trading. That makes Calmar the closest of the common ratios to a prop firm's own test. It still has three gaps:
- It does not know your limit. As a hypothetical, not any firm's rule: if an account is closed at a 3% drawdown, trader B's 3.28% would have ended it. B's Calmar of 22.3 still looks healthy. Compare max drawdown with your actual limit first, and use Calmar to rank the accounts that survive.
- It uses end-of-day closes. ExpectancyIQ builds the drawdown from daily returns, so a dip that recovers before the session ends does not show. Firms that track equity in real time can breach on a move that Calmar never sees. The difference is covered in end-of-day vs intraday drawdown.
- Firms measure the floor differently. Some trail a floor up from your highest balance, and some use a static limit from the starting balance. Calmar's peak-to-trough percentage matches neither exactly. The prop firm drawdown calculator applies a firm's own rule, and the current per-firm rules are listed separately because amounts change.
How many days before the Calmar ratio means anything
Calmar rests on a single observation, the worst drawdown, so it is the noisiest of the common ratios on a short history. Add one more day to trader A: an eleventh day of −1.5% lifts the max drawdown from 0.90% to 1.50% and cuts the average return to 0.127% a day. The Calmar falls from 81.2 to 21.4 on one day's data.
Two properties follow. The maximum drawdown can only stay the same or grow as history lengthens, so a young account's Calmar can fall sharply when it meets its first real losing streak. And an account that has only gone up has no drawdown at all, so there is no ratio to show. Wait for a drawdown you would be unhappy to repeat, and at least a few months of days, before comparing. The sample-size problem is the same as for win rate; see how many trades before you can trust a statistic.
What is a good Calmar ratio?
There is no official threshold. The "above 3 is good" rules of thumb online do not trace to a primary source I could check, so treat them with suspicion. What a number means depends on:
- The window. A 36-month Calmar, the full-history MAR ratio and a three-month figure from the same account can all differ widely.
- How the return is annualised. Compounding trader A's ten days over 252 gives about 106% a year instead of 73.08%, which makes the Calmar 118 rather than 81. Pick one method and keep it.
- The size of the drawdown. 0.5% and 5% are different risks even at the same ratio. A Calmar of 3 with a 25% drawdown is a very different account from a Calmar of 3 with a 2% drawdown.
Track your own Calmar by quarter, using one window and one return method, and look at the size of the drawdown beside it.
Where the Calmar ratio misleads traders
- One bad stretch sets the denominator. A single losing streak controls the whole ratio, and it can be an old one that your current rules would have avoided.
- It cannot separate edge from luck. Scaling position size scales the return and the drawdown together, so the ratio barely moves, and a lucky run scores as well as a real edge. Use expectancy and profit factor to judge the edge itself.
- Closing prices hide the real low. Daily data misses an intraday dip, as covered above.
- It says nothing about how often you lose. The gain to pain ratio counts every loss, not only the deepest stretch.
Calmar ratio vs the other trading ratios
- Sharpe ratio: return against all volatility; see the Sharpe ratio guide.
- Sortino ratio: return against downside deviation only; see the Sortino ratio guide.
- Recovery factor: net profit in dollars divided by the largest dollar drawdown. It uses total profit, where Calmar uses the annualised rate.
- Expectancy and profit factor: measure the average and the balance of your trades, not the path of your account.
Your own Calmar ratio, without the spreadsheet
ExpectancyIQ's Performance report shows an annualised Calmar ratio next to Sharpe and Sortino, using the formula above over every trading day in the range, and a dash when there is no drawdown yet. A trading day follows the 17:00 CT session roll described in the methodology. The trading dashboard shows how it sits with the rest of the numbers, and the glossary defines each metric in short.