Risk
What is a drawdown, and why every trading system has one
Every trading system, without exception, has losing stretches. The polite word for them is drawdown, and understanding the number is the difference between sitting through a bad month and panicking out at the bottom of one.
The definition
A drawdown is how far an account has fallen from its highest point so far, expressed as a percentage of that high.
The high is the important half. A drawdown is not measured from what you deposited. It is measured from the best your account has ever been, which moves up every time you set a new record.
An example. You deposit $2,000:
| Balance | Highest so far | Drawdown | |
|---|---|---|---|
| Start | $2,000 | $2,000 | 0% |
| After a good run | $3,000 | $3,000 | 0% |
| After a bad run | $2,550 | $3,000 | −15% |
| Recovering | $2,850 | $3,000 | −5% |
| New high | $3,200 | $3,200 | 0% |
At $2,550 you are still $550 up on your deposit. You are also in a 15% drawdown. Both are true, and the second one is what it feels like.
The largest of these figures over a period is the maximum drawdown, and it is the single most useful number for deciding whether you could actually live with a system.
Why the maths of recovery is unfair
Here is the part that surprises people. Getting back to even takes a bigger gain than the loss that put you there.
If you lose 20%, you need 25% to recover. Not 20%.
The reason is that the gain is calculated on a smaller balance. Lose 20% of $1,000 and you have $800. To get back to $1,000 from $800 you need $200, and $200 is 25% of $800.
| Drawdown | Gain needed to recover |
|---|---|
| −5% | +5.3% |
| −10% | +11.1% |
| −20% | +25% |
| −30% | +42.9% |
| −50% | +100% |
| −70% | +233% |
| −90% | +900% |
The curve is gentle at the top and vicious at the bottom. A 10% fall is an ordinary inconvenience. A 50% fall means doubling your money just to get back where you started.
This is the real argument for controlling position size, and it is why a system that risks a fixed small slice of the balance per trade behaves so differently from one that raises its stake to chase a loss.
The two questions that matter
When you look at any record, ask two things.
How deep? The maximum drawdown tells you the worst it got. Then ask yourself, honestly, whether you would still have been holding at that point. Most people overestimate themselves here. A 16% fall reads as a small number on a page and feels quite different when it is your money and you do not know whether it stops there.
How long? Depth is only half of it. A 15% fall that recovers in three weeks is a different experience from a 15% fall that grinds on for nine months. The second one is where people give up, usually shortly before it turns.
Our own worst stretch
Goldsmith's deepest fall over the recorded eleven months was 16.7%, in February 2026. The account sat 16.7% below its previous high before recovering.
On the smaller account sizes the figure was slightly shallower, because the broker's 0.01 lot minimum rounds every trade down and so carries a little less risk than intended:
| Deposit | Deepest fall | In dollars |
|---|---|---|
| $1,000 | −14.2% | −$202 |
| $2,000 | −14.1% | −$522 |
| $5,000 | −16.3% | −$1,698 |
| $10,000 | −16.5% | −$3,471 |
| $20,000 | −16.7% | — |
How Goldsmith would have performed on Vantage's recorded prices, 1 Nov 2025 – 23 Sep 2026. Not a live account. Past performance does not guarantee future results, and you can lose some or all of your deposit.
There was also one losing month in the eleven, December 2025, at −3.7%. On a $2,000 account that was −$66.
Look at the dollar column rather than the percentage, because that is what you would actually have watched. On a $2,000 account, the worst stretch was $522 disappearing across a couple of weeks in February. If that number makes you uncomfortable, that is useful information, and it is better learned here than in the middle of it.
A future drawdown could be deeper than any of these. A recorded maximum is the worst that happened in one particular stretch of market history. It is not a limit, and nothing prevents a worse one.
Why a drawdown is not a malfunction
There is a tempting thought during a losing run: something has broken, the market has changed, the system no longer works.
Sometimes that is true. Far more often it is just how probability looks close up. Any system that wins less than 100% of the time will string losses together sometimes, and the longer you run it the more certain that becomes. If a system has never had a losing month, the usual explanation is that it has not run for very long.
What separates a normal drawdown from a real problem is not how it feels. It is whether the behaviour still matches what the rules say it should do. That is a question about the rules, not about the balance.
What to do about it
Nothing here tells you to buy, sell or hold anything. But three things are worth settling before you ever see a drawdown, rather than during one:
Decide what you could sit through. Not what sounds acceptable. What you would actually hold through without touching anything. Then size the account so the recorded worst case lands inside it, with room to spare, because the next one may be deeper.
Use money you can afford to put at risk. Money you need within a year does not belong in a leveraged gold account. The pressure changes your decisions at exactly the wrong moment.
Judge over a sensible stretch. Eleven months is a reasonable sample. Eleven days is noise. Most people who abandon a system do it at the bottom of a drawdown, which is the one moment when the decision is hardest to reverse.
In short
- A drawdown is the fall from an account's highest point, not from your deposit.
- Recovery needs a bigger percentage than the fall, and the gap widens fast below 30%.
- Depth and duration both matter. Duration is what makes people quit.
- Goldsmith's deepest recorded fall was 16.7%, in February 2026. A future one could be worse.
- A losing run is not evidence that something is broken. It is the normal shape of trading.
The complete month-by-month record, including every losing month, is on the results page.