Insights · Risk Management
Maximum drawdown is not enough
Published: · by the QuantGovernor team
Maximum drawdown — the worst peak-to-valley loss in an account's history — is the most quoted risk number in trading, and on its own it is not enough to judge a strategy's risk. It is one snapshot of one realized outcome. Real risk lives in several dimensions at once: how the drawdown is measured, what exposure was open while the curve looked calm, how positions relate to each other, and what mechanism — if any — caps the damage. This article walks through each.
Which drawdown are you looking at?
The first question is almost never asked: drawdown of what? Balance drawdown counts only closed trades — it is what the account has admitted to. Equity drawdown includes the floating profit and loss of open positions — it is what the account actually lived through. For a strategy that closes every trade with a stop, the two are close. For a strategy that holds losing positions open — grids, martingales, "it will come back" trading — balance drawdown can look serene while equity drawdown tells a horror story. If a provider quotes one number, ask which one it is. We covered how to see this on a public track record in How to read a Myfxbook page.
Realized vs unrealized: the loss that hasn't confessed yet
A loss does not start existing when a position is closed — it exists the moment the market moves against an open position. Keeping a losing trade open converts a visible realized loss into an invisible unrealized one, at the price of growing exposure. That is why floating P&L is not an accounting detail: it is where risk hides. When we declare that a component of a system uses position averaging, as with part of Copy Gold, the honest consequence is exactly this: its equity line deserves more attention than its balance line, and we say so on the page.
Exposure and concentration: risk before the loss
Drawdown measures damage after it happened. Maximum exposure — how much capital was committed to open positions, with how much leverage, at the worst moment — measures the damage that could have happened. Two accounts with an identical 10% historical drawdown are not equally risky if one reached it holding a single position with a tight stop and the other was holding ten averaged positions on the same instrument at 1:100 leverage. Concentration matters too: five positions on five truly different markets are not the same animal as five positions that are all, in substance, a bet on the same thing.
Correlation: the multiplier nobody prices in
Positions that lose together, draw down together. Correlation between instruments — and between strategies — decides whether a portfolio's risks add up or offset. The trap is that correlation is not constant: in calm markets, strategies can look pleasantly independent; in stress events, correlations tend to rise exactly when you need them low. This is why a multi-strategy portfolio is only as diversified as its behavior in bad weeks, not its average — a topic big enough that we treat it separately in a dedicated article on correlation risk.
Leverage: the amplifier
Leverage does not create risk out of nothing — it multiplies whatever risk profile already exists. A modest strategy at high leverage and an aggressive strategy at low leverage can produce the same drawdown with completely different failure modes: the leveraged one is exposed to gaps and margin calls, where losses can exceed any historical statistic. Any risk discussion that ignores the leverage actually used is incomplete.
The stop logic: what actually caps the damage
Finally, the question that separates risk management from risk description: what mechanically stops the loss? There are three honest answers: a stop loss on every trade, a portfolio-level equity stop, or nothing but hope. Our systems use the first two — XAUUSD Adaptive Breakout defines a stop and target on every single trade, and Olympus adds a declared −30% equity stop over the whole portfolio. And one caveat we repeat everywhere: declared limits are risk-management targets, not guarantees — in extreme conditions (gaps, slippage, illiquidity) they can be exceeded. Anyone who tells you otherwise is selling certainty that does not exist.
The five questions that complete the picture
- Is the quoted drawdown measured on balance or on equity?
- What was the maximum open exposure, and at what leverage?
- How concentrated is the book — how many truly independent bets?
- How do the components behave in stress, not on average?
- What mechanism, exactly, caps the loss — and what are its limits?
Ask these of any strategy, ours included. A provider who answers all five without flinching is telling you about risk; one who quotes a single flattering number is telling you about marketing.
Terms used in this article, explained in the glossary: drawdown · equity stop · slippage · averaging · leverage · correlation.
Related: How to read a Myfxbook track record · Risk Disclosure. Past performance is not indicative of future results.