Drawdown is what decides whether a system is tradeable
A profitable system with an untradeable drawdown is not a system. It is a spreadsheet result. The most common reason a strategy that backtests well fails in live trading is not that the edge disappeared — it is that the drawdown arrived, and the trader either reduced size at the wrong moment, abandoned the rules, or was simply forced out by margin.
So the useful question is not “what does this make?” but “what does it make, and what does it put me through on the way?”
Drawdown at any point = Running peak − Current equity
Recovery return = Drawdown % ÷ (100 − Drawdown %)
The asymmetry that makes drawdown expensive
Gains and losses do not compound symmetrically, because they act on different bases. A 20 percent loss leaves you needing 25 percent to recover. A 33 percent loss needs 50 percent. A 50 percent loss needs 100 percent. A 75 percent loss needs 300 percent.
Plot that relationship and the practical conclusion is hard to miss: the cost of a drawdown grows faster than the drawdown itself, so a strategy that routinely runs to 40 percent is not twice as risky as one that runs to 20 percent. It is closer to four times as risky, because the recovery hurdle and the probability of abandonment both scale superlinearly.
Why drawdown scales with position size
For a fixed sequence of trades expressed in R, percentage drawdown scales roughly linearly with the amount risked per trade. This is one of the few genuinely useful levers in system design, because it lets you keep the same edge and simply change the depth of the hole.
- A sequence producing 10R of peak-to-trough loss is a 20 percent drawdown at 2 percent risk per trade.
- The same sequence is a 10 percent drawdown at 1 percent risk.
- And a 40 percent drawdown at 4 percent risk.
The edge does not change. The experience does. Given that most traders overestimate their tolerance for losses, sizing to a drawdown you can actually sit through is worth more than optimising the entry rule.
Why losing streaks matter more than the drawdown number
Drawdown is an abstraction. A streak is not. If a system's worst historical run is eight consecutive losses, then the next time it prints five losses in a row you will know that four more are within the historical range, and the temptation to intervene on trade six is the main way systems get broken.
The mathematical expectation for streak length is often longer than people assume. At a 40 percent win rate, the probability of five consecutive losses is about 7.8 percent per starting point, and over a few hundred trades a streak of eight or nine is unremarkable. Systems are abandoned for behaving normally.
What this calculator does not do
It measures the sample you paste into it. If the sample is one favourable period, the drawdown it reports is a floor, not a forecast. Drawdown is path-dependent: a different ordering of the same trades produces a different maximum, and the ordering you get live is not the ordering you tested.
For that reason, a useful discipline is to take a trade sequence, reshuffle it many times, and look at the distribution of drawdowns rather than the single figure from the original order. The original order is one draw from a distribution, and it is not a particularly informative one.
Reading the output
- Maximum drawdown in percent is the headline. Compare it to what you can hold through, not to what other systems report.
- Recovery return is the hurdle. If it is large, remember that the system must clear it before it produces a single unit of new profit.
- Longest losing streak is the thing you will actually feel. If it is longer than you can tolerate, reduce size rather than the streak.
- Largest single loss should be close to your intended per-trade risk. If it is much larger, slippage, a gap or a missed stop is in the record and should be investigated.