Maximum drawdown is the largest fall your equity curve has taken between a peak and the trough that followed it. It is the measure that answers the one question P&L never asks: how much were you down, at the worst moment, before recovering?
From peak to trough
Contents
Drawdown is the gap between a peak in your equity curve and the lowest point reached after that peak, before a new peak is made. Maximum drawdown is simply the largest of all those found over the period. The words 'after that peak' are not a detail: they are what makes the measure.
The difference from an ordinary loss lies in that reset. As soon as a new peak is reached, the counter restarts, because at that moment you are no longer behind yourself. That is why an account can end the year well up with a maximum drawdown of 30%: the two figures describe different things.
Its real function is psychological and operational more than statistical. A 40% drawdown does not read as a number but as four months of losing, during which the same method has to keep being executed. That is where most profitable systems get abandoned, and that is what this figure is for anticipating.
Maximum drawdown = max over the period of (peak − following trough) ÷ peak
The calculation runs on the equity curve, so on results cumulated in order. Changing the order of the same trades changes the drawdown, while the final P&L does not move.
It exists in balance (closed trades) and in equity (open positions included). The second version is always harsher, and it is the one prop firms watch.
The duration of the drawdown does not appear in the formula, and yet it is what makes people quit: two 20% falls, one over three days and one over seven months, give the same figure.
| Maximum drawdown | What it implies |
|---|---|
| under 10% | Comfortable. Compatible with most prop firm rules without thinking about it. |
| 10 to 20% | Normal for discretionary trading. That is already several weeks of doubt to get through. |
| 20 to 30% | High. It takes +43% to come back from a 30% fall, and the method is rarely held that far. |
| above 30% | The system can be profitable and still unliveable. This is the zone where people switch method at the worst moment. |
| above 50% | You have to double the remaining capital to break even. Almost always a position sizing error. |
The asymmetry in the table is the point to remember, and it is arithmetic: a 30% fall requires a 43% rise to recover, and a 50% fall requires 100%. The deeper the drawdown, the more disproportionate the recovery, which is why limiting the fall matters more than maximising the rise. That asymmetry is invisible in an annual P&L.
The second point is that the figure says nothing about duration. A 20% fall crossed in three days looks like an accident; the same one spread over seven months is a daily crisis of confidence. It is the second that makes people abandon methods, and the formula confuses them completely.
Maximum drawdown is 15.5%, and it is computed from the $11,600 peak, not from the starting capital. That is the point everybody misses: measured from $10,000, the $9,800 trough would be just 2%, and that figure would describe nothing of what was lived through.
The maximum drawdown of a track record gives the order of magnitude of what you have to be ready to take. Halving position size roughly halves the expected drawdown, and it is the one lever that acts directly on it without touching the method.
It is also the calculation to run before a prop firm challenge. If your method's historical drawdown exceeds the challenge's limit, failure is not a question of discipline, it is a question of arithmetic.
Two systems with the same annual return and drawdowns of 8% and 35% do not ask for the same person. The second is perfectly profitable on paper and will be abandoned in the third month of falling, which makes it worthless in practice.
Looking at this figure before choosing a method avoids discovering mid-course that you cannot follow it, and that discovery always happens at the worst moment.
The first flaw is measuring the fall from starting capital instead of from the peak. It is the most frequent mistake and it systematically minimises the figure, sometimes by a factor of five as in the example above. An account that rose and then fell back lived through a decline that is invisible if measured from the start.
The second is duration missing from the formula. Maximum drawdown is a depth figure, and what makes people abandon a method is length. Both can be measured, only one is displayed everywhere, and that is why a 15% drawdown lived over six months always surprises those who only looked at the percentage.
The third is reading it without the trade count. A maximum drawdown measured over forty trades says almost nothing: the worst has probably not happened yet, and the figure displayed is an optimistic minimum rather than an estimate. On a short history it is a lower bound, not a forecast.
The classic mistake. Measuring the fall from starting capital. In this page's example, the $9,800 trough gives 15.5% measured from the peak and only 2% measured from the initial $10,000. The second figure is reassuring and describes nothing of what the trader actually went through.
| Indicator | Formula | What it measures | When to use it |
|---|---|---|---|
| Maximum drawdown | (peak − following trough) ÷ peak | The worst decline lived through, in % | Sizing positions, choosing a method |
| Recovery factor | net P&L ÷ maximum drawdown | What each unit of decline endured returns | Comparing two systems at equal risk |
| Trailing drawdown | A limit that follows your high | The rule specific to prop firms | Surviving a challenge, not judging a system |
| Sharpe ratio | mean ÷ standard deviation of results | Steadiness, not depth | Judging the stability of a performance |
It is the only lever that acts directly and proportionally. Halving size roughly halves drawdown, with no entry rule touched. Many traders rebuild their strategy when their problem lay entirely in sizing.
Add next to the percentage the number of days spent below the peak. That is the figure predicting abandonment, and it appears in almost no tool. Recording it once a quarter is enough to know whether your method is holdable, and not merely profitable.
A drawdown is rarely spread evenly across your setups. Look at each one's drawdown separately: sometimes a barely profitable configuration is responsible for most of the hole, and stopping it improves the curve far more than optimising the one that already works.
It is the largest fall in your equity curve between a peak and the trough that followed it, expressed in currency and as a percentage. The essential point is that it is measured from the peak, never from your starting capital, and that a new peak resets the counter.
Under 10% is comfortable, between 10 and 20% is normal for discretionary trading, and above 30% a system can be profitable while staying unliveable. The reason is arithmetic: recovering from a 30% fall requires a 43% rise, and few traders hold the same method that long.
By reducing position size, first of all. It is the only lever that acts almost proportionally and requires changing no entry rule. The second lever is stopping the configuration that digs the deepest, which is not always the one that earns least.
Maximum drawdown is a measure of your history: it describes what your system put you through. Trailing drawdown is a prop firm rule: a loss limit that follows your high, so it rises as you win. An account in profit can be disqualified by the second while the first shows nothing alarming.
No. Over forty or fifty trades it gives the worst of what has happened so far, which is not the same as the worst that can happen. Read it as a lower bound: the real question is whether you would hold through twice as much, because that is what will eventually occur.
Tradoshi computes maximum drawdown from each peak of your equity curve, not from your starting balance, which is exactly the distinction that makes all the difference in the example above. It appears in currency and in percentage in the Reports screen, next to profit factor and average R, and it recomputes on every sync with your broker.
The Prop firm screen handles the other question, that of funding rules, which is not the same: a static drawdown is measured from your starting balance, a trailing drawdown follows your high. The matching glossary term explains why an account in profit can be disqualified by the second.
To go deeper on this: read the full guide. Every term is gathered in the glossary.