Understanding drawdown: what the number means and what it hides

Drawdown is the most quoted and least understood number in automated trading. Every product listing and every backtest report shows one, and almost every buyer reads it as a promise: “the most I can lose is 18%”. It is not. Drawdown is a measurement of a specific path that prices took in the past, and the future will take a different one. This guide explains what the different drawdown figures mean, why they understate risk, and how to use them sensibly when sizing an account for an Expert Advisor.

Definitions

MetaTrader reports and third-party verification services use several related terms:

• Absolute drawdown — how far the balance fell below the initial deposit at its worst point. If you started with 10,000 and the lowest balance was 9,400, absolute drawdown is 600. It says nothing about later losses from a higher peak.
• Maximal drawdown — the largest peak-to-trough fall in balance during the test, in currency and as a percentage of the peak. This is the “max DD” figure most people quote.
• Relative drawdown — the largest percentage fall from any peak, which can be a different event from the largest currency fall when the account has grown.
• Floating (equity) drawdown — the fall in equity, which includes open, unrealised losses. Balance drawdown only counts closed trades. A grid or averaging EA can show a smooth balance curve while equity is deeply underwater; floating drawdown exposes that.

When a product page on this site shows a drawdown figure, it labels which kind it is and whether the source is a backtest, a vendor claim or a live record. If a figure has no label, treat it as balance drawdown from a backtest and assume floating drawdown was worse.

Why balance drawdown flatters averaging strategies

Strategies that add to losing positions — grids, martingale variants, “recovery” modes — rarely close trades at a loss. Their balance curve therefore looks almost straight, and maximal balance drawdown can be tiny. Meanwhile the equity curve records every open loss, and during a strong trend the floating drawdown can exceed the balance drawdown by a factor of ten or more. If a listing shows a 3% drawdown for a strategy that never closes losers, ask for the equity drawdown; if it is unavailable, the number you have is not informative. Our guide Reading backtests critically shows how to reveal this from the tester's equity chart.

Historical drawdown is a floor, not a ceiling

A backtest covers one realisation of history. The maximal drawdown it reports is simply the worst run of losses that happened to occur in that sample. Three things guarantee the future figure will be larger at some point:

1. Sample size. The longer a strategy runs, the more opportunities it has to hit an unusual streak. A drawdown measured over 200 trades is systematically smaller than the drawdown the same strategy will experience over 2,000.
2. Regime change. Volatility, spreads, session behaviour and correlations shift. A strategy tuned to one regime meets its worst drawdown when the regime changes — which is, by construction, not in the tuning sample.
3. Execution. Live slippage, requotes and widened spreads at news add losses that the tester models imperfectly, and they cluster in exactly the periods when drawdowns are already forming.

A common rule of thumb is to plan for at least one and a half to two times the historical maximal drawdown. That is not a statistical guarantee either; it is an acknowledgement that the report is optimistic.

Drawdown duration matters as much as depth

Two strategies with the same 20% maximal drawdown can feel completely different. One spends a week underwater and recovers; the other spends nine months. Reports show this as “longest drawdown period” or can be read from the equity chart. Long, shallow drawdowns are the ones that make traders switch off a tool a week before it recovers, which is how a strategy that was profitable in the long run loses money for its user. Before running a tool, decide in advance how long a losing period you are prepared to sit through and confirm that the historical record contains periods that long.

Sizing an account around drawdown

The arithmetic is simple. Suppose a tool's report shows a 22% maximal balance drawdown at a fixed lot size of 0.10 on a 10,000 account, and you plan for twice that: 44%. If you cannot tolerate more than a 20% loss on your own account, you need either a smaller lot size per unit of balance or a larger balance for the same lots. Halving the lot size roughly halves the drawdown in currency, so 0.05 lots on the same 10,000 brings the planning figure to about 22% — inside your limit.

Most tools sold here offer risk-based sizing, where the lot is computed from a percentage of balance per trade. That keeps the percentage drawdown roughly constant as the account grows or shrinks, which is what you want. The risk management controls guide explains the sizing inputs and the daily and equity stops that cap losses when a drawdown runs faster than expected.

Drawdown and leverage

Leverage does not change a strategy's drawdown in currency; it changes how much margin you need to hold the positions. Where it bites is at the extreme: during a deep floating drawdown the broker may issue a margin call and close positions at the worst possible moment, turning a temporary equity drawdown into a permanent balance loss. Check the product page for the minimum recommended balance and the “insufficient margin” section of our troubleshooting guide. If a tool's floating drawdown would trigger a margin call at your leverage and balance, your account is too small for that tool.

Drawdown under prop-firm rules

Prop firms measure drawdown in their own ways — often a daily equity drawdown from the day's starting balance and an overall trailing drawdown — and breaching either ends the account. A tool with a comfortable 15% historical maximal drawdown can still fail a 5% daily limit on a single bad day. Compare the tool's worst daily equity swing, not just its maximal drawdown, against the firm's rules; Prop-firm compatibility covers the details.

Reading drawdown on this site

Every performance record on a product page carries a source label: vendor-provided, backtest, live-verified or demo. Sample listings created to demonstrate the site show figures marked “Sample data” which mean nothing at all. For genuine records, look for the equity drawdown, the test period and trade count, and the sizing used. If the sizing is not stated, the drawdown percentage cannot be compared with anything.

Summary

• Know which drawdown you are looking at: balance, equity, absolute, maximal or relative.
• For strategies that hold losing positions, only equity drawdown is meaningful.
• Treat historical drawdown as a floor; plan for one and a half to two times the figure.
• Consider duration as well as depth, and decide your tolerance before starting.
• Size lots so the planning figure fits within your tolerance and never approaches a margin call.

Drawdown is a useful number precisely when you stop treating it as a limit and start treating it as the beginning of a conversation about risk. None of this is investment advice; it is arithmetic you should do before any tool touches a live account.