Risk and Capital
Ask a room of retail traders what an acceptable drawdown is and you will hear numbers between five and fifty per cent, usually delivered with confidence and no reasoning. Ask a risk manager and you will get a question back: drawdown relative to what, measured over what horizon, produced by which strategy?
Three drawdowns that mean different things
01
Statistical drawdown
The losing run your edge produces naturally. Predictable in shape, unpredictable in timing. Costs money, tells you nothing.
02
Regime drawdown
The market changed and your edge no longer fits it. Looks like bad luck for weeks before it looks like a signal.
03
Behavioural drawdown
You stopped following your own rules. The only kind that is entirely within your control, and the most common.
The whole discipline of drawdown management is telling these apart quickly enough to act. A journal that records rule adherence separately from outcome is the only reliable instrument for doing it.
What the numbers usually look like
For a discretionary intraday FX book risking half a per cent per idea, a peak-to-trough drawdown of six to ten per cent inside a year is unremarkable. For a slower swing approach with a lower win rate and a higher payoff ratio, twelve to eighteen per cent is within normal range. Neither number is a target; both are the price of the return distribution you chose.
Expected versus tolerable
Expected drawdown comes from your track record and your risk per trade. Tolerable drawdown comes from your balance sheet and your nervous system. Professionals size so the first is comfortably inside the second, and they write both numbers down before the year starts.
- Estimate the worst losing run your win rate makes likely across a year.
- Multiply by your risk per trade to get expected drawdown.
- Write down the loss at which you would genuinely change behaviour — that is tolerable drawdown.
- If expected is within a few points of tolerable, reduce risk per trade until it is not.
The recovery arithmetic nobody likes
Losses and gains are not symmetric. A twenty per cent drawdown requires a twenty-five per cent gain to get back. A fifty per cent drawdown requires a hundred. This is the entire argument for small position sizes stated in one line, and it is why professionals spend far more time on the downside than on entries.
You cannot compound a return you did not keep. Capital preservation is not conservatism, it is the mechanism.
When drawdown becomes information
Treat a drawdown as information when three things happen together: the losses are clustered in one setup, the market structure you rely on has visibly changed, and your rule adherence score is high. High adherence plus poor results is the clearest evidence you have that the edge, not the trader, is the problem.
Written by Adrian Vestberg — former interbank FX dealer, now running a two-person discretionary desk.