Part 1: Risk and its numbers · Chapter 4 of 14 · 12 min
The drawdown
The drawdown is the distance between the account and its previous high. In our series the maximum fall is 35.3 R (levels bounce), 18.9 R (swing break), 1.7 R (Asia hours) and 9.0 R even in the example with a true edge; the series with an edge spent 64% of trades below its own high, those without a proven edge from 84% to 96%. To recover a −50% you need +100%. With 2% risk, 20 R of drawdown is −33% of the account.
What the drawdown measures
The drawdown is the distance between the current capital and the high reached before. If the account reached 11,000 $ and is now at 9,900 $, the drawdown is 1,100 $, 10% of the high. It is measured in three ways, and we look at all of them:
- in R: the sum of results from the high downwards (the measure that does not depend on how much you risk);
- in percent of capital: it depends on the risk per trade (we convert it further on);
- in duration: how much time (trades or days) passes between one high and the next, that is how long you stay "underwater".
The maximum drawdown is the deepest one seen in the whole series. It is a record: it grows every time a period worse than all the previous ones arrives, and it grows even for a system that works.
The real series
Let us measure the maximum drawdown in the four series, starting from the sum of R in time order:
| Series | Maximum drawdown | From the peak | To the low | Final sum | Trades below the high | Longest period below the high (closed) |
|---|---|---|---|---|---|---|
| LA | 35.3 R | 13 May 2026 (+17.0 R) | 4 October 2026 | −18.2 R | 96% | 91 trades, 55 days |
| LD | 18.9 R | 18 August 2026 (+23.8 R) | 18 September 2026 | +10.6 R | 92% | 41 trades, 39 days |
| SD | 1.7 R | 20 March 2026 (after 18 trades) | 12 May 2026 | +5.1 R | 84% | 16 trades, 90 days |
| SX | 9.0 R | 13 August 2026 (after 441 trades) | 20 August 2026 | +114.0 R | 64% | 60 trades, 28 days |
Let us read the table.
- LA reached +17.0 R on 13 May and has not touched the high since: from that day to 4 October it is 282 trades and 144 days, with the account 35.3 R below the peak. A trader risking 2% would have seen the account fall 53%.
- LD earned +10.6 R in total but had a drawdown of 18.9 R in a single month (18 August-18 September). Whoever looks only at the final result does not see that on 18 August it stood at +23.8 R.
- SD has a small drawdown in R (1.7 R) because its losses are small (0.35 R on average), but it stayed below the high for 90 days: depth does not tell duration.
- SX, with a built-in edge, has a maximum drawdown of 9.0 R against a final gain of 114 (ratio 12.7), and spent 64% of trades below the high. The first figure suggests an easy system, the second recalls that even with a true edge the account is almost always below its high.
Being below the high is the normal condition: the series is at its high only in the few trades where it sets a new record. A system that is almost always below its high is not a broken system.
Read the chart in words
Daily gold chart with, below, the cumulative R curve of LA and the bars of the distance from the previous high.
Read the chart in words
Daily gold chart with, below, the cumulative R curve of the synthetic example and the bars of the distance from the previous high.
Recovery is not symmetric
A percentage loss is recovered with a bigger percentage gain, because you start from a smaller capital:
| Loss | 10% | 20% | 30% | 40% | 50% | 60% | 75% | 90% |
|---|---|---|---|---|---|---|---|---|
| Gain to recover | +11.1% | +25.0% | +42.9% | +66.7% | +100% | +150% | +300% | +900% |
If the account falls 50%, you have to double it to get back where you were. Beyond a 50% loss the road becomes unrealistic for a system that earns a few percentage points a year. It is the main reason why the risk per trade is chosen by looking at the drawdown and not at the expected gain.
How much the maximum drawdown grows
The maximum drawdown is not a fixed number of the system: it depends on how many trades you have made. With the bootstrap (chapter 2) I draw the results of each series at random and measure the maximum drawdown of paths of 100, 250 and 500 trades. Median and, in brackets, the 5th to 95th percentile range, in R:
| Series | 100 trades | 250 trades | 500 trades |
|---|---|---|---|
| LA (mean −0.03) | 12.0 (5.9-24.0) | 21.2 (10.0-40.6) | 33.2 (15.9-62.6) |
| LD (mean +0.03) | 9.4 (5.1-19.4) | 14.5 (7.8-29.3) | 19.8 (11.0-39.5) |
| SD (mean +0.09, 55 trades) | 3.3 (1.8-6.5) | 4.5 (2.8-8.1) | 5.4 (3.6-9.2) |
| SX (mean +0.21, true edge) | 6.0 (3.0-11.0) | 8.0 (5.0-13.0) | 9.0 (6.0-15.0) |
For a system without an edge (LA, LD) the maximum drawdown grows fast: from 100 to 500 trades, for LA, it goes from 12 to 33 R. With a true edge (SX) it grows much more slowly (from 6 to 9 R), because the account rises and every new high resets the distance. For SD, where losses are small, the drawdown is low in absolute terms, but it depends on only 55 observed trades.
The practical result: whoever looks at the maximum drawdown of the first months of a system is looking at a number that will grow. With 250 trades, the median of LD is 14.5 R; with 500 trades it is 20 R, and one time in twenty it exceeds 39 R.
From drawdown in R to drawdown in percent
With a constant risk in percent of current capital, the drawdown in percent depends on how many R you lose and how much you risk. For small percentages it is about "drawdown in R × risk"; for large drawdowns it is computed with 1 − e^(−risk × R), because each loss is computed on the capital that remains:
| Drawdown in R | Risk 0.5% | Risk 1% | Risk 2% | Risk 5% |
|---|---|---|---|---|
| 5 R | 2.5% | 4.9% | 9.5% | 22.1% |
| 10 R | 4.9% | 9.5% | 18.1% | 39.3% |
| 15 R | 7.2% | 13.9% | 25.9% | 52.8% |
| 20 R | 9.5% | 18.1% | 33.0% | 63.2% |
| 30 R | 13.9% | 25.9% | 45.1% | 77.7% |
| 40 R | 18.1% | 33.0% | 55.1% | 86.5% |
With this table and the bootstrap one you answer the question that matters: "with my system and my risk, how much can I lose in a year of trading?" For paths of 500 trades made by drawing from the results of the series, the maximum drawdown in percent (median, with the 5th to 95th percentile in brackets) is:
| Series | Risk 1% | Risk 2% | Risk 5% |
|---|---|---|---|
| SX (true edge) | 8.7% (5.9-14.1) | 17.0% (11.6-26.7) | 38.7% (27.4-56.4) |
| LD (mean close to zero) | 18.5% (10.8-33.2) | 34.6% (20.8-56.7) | 69.6% (46.5-90.6) |
And on the real series in the order in which they happened, with constant risk on the current capital:
| Series | Risk 1%: final / drawdown | Risk 2%: final / drawdown | Risk 5%: final / drawdown |
|---|---|---|---|
| LA | 81% / 30.7% | 62% / 53.3% | 20% / 87.9% |
| LD | 109% / 17.4% | 115% / 32.2% | 108% / 63.7% |
| SD | 105% / 1.7% | 110% / 3.4% | 126% / 8.3% |
| SX | 304% / 8.7% | 876% / 16.7% | 15,049% / 37.4% |
(Final capital as a percentage of the initial one.) With 5% risk LD ends at +8% but goes through a −63.7%: a result nobody bears. SX at 5% reaches 15,049% because compound interest works on an edge built by hand, repeated for 552 trades: it shows how far capital can grow with a true edge, and promises nothing.
What to do with it
- Decide beforehand the maximum drawdown you are willing to bear (in percent of the account), not after seeing it.
- Estimate it with your system: bootstrap on your results, drawdown at 250-500 trades, 95th percentile.
- Choose the risk per trade so that this 95th percentile stays under your limit.
- Look at the duration too: a drawdown of 55-90 days below the high is normal and must be allowed for.
- Do not confuse a drawdown within the expected one with a broken system, nor a small drawdown with the certainty that the risk can be raised.
Procedure in four steps
- Record the account trade by trade, with the high reached.
- Compute the current drawdown, the maximum and the period below the high.
- With the bootstrap on your results estimate the drawdown at 500 trades (95th percentile).
- Choose the risk per trade so that this value, turned into a percentage, stays within the limit you set.
Exercises
Chapter exercises
1. The account has fallen 40% from its high. What percentage gain is needed to return to the high?
2. With 2% risk per trade, roughly what percentage of the account is a drawdown of 20 R?
3. In LA the account touched +17.0 R on 13 May 2026 and in October stood at −18.2 R. What is the drawdown?
4. A system has a maximum drawdown of 9 R after 100 trades. What should you expect with 500?
Have a question or want to share your exercise?
Post in the community, or join the free signals room on Telegram.
Educational content, not financial advice. Trading involves risk.