Almost every guide to forex risk management says the same four things: risk one percent, use a stop-loss, size your positions, mind your risk-reward ratio. All of it is correct, and none of it explains why traders who follow the rules still lose more than their spreadsheet predicted. This guide covers the rules briefly, then spends most of its length on the gap between a risk plan and what the market actually does to it.
The rules, stated once
Risk a fixed small fraction per trade. One to two percent of account equity is the conventional range. The point is not the specific number; it is that a run of losses is survivable arithmetic rather than an account-ending event.
Define the exit before the entry. A stop-loss decided after a position moves against you is not a risk control, it is a negotiation you will lose.
Size the position from the stop, not the other way round. Distance to the stop and the amount you are willing to lose together determine the position size. Choosing a lot size first and placing the stop wherever it lands is the most common way a disciplined-looking plan produces undisciplined risk.
Require the reward to justify the risk. A strategy winning less than half its trades can be profitable if the winners are larger. A strategy risking three to make one needs to be right almost always, which nothing is.
That is the standard advice, and it is genuinely the foundation. What follows is what it does not cover.
Where a risk plan meets execution
Your stop is an instruction, not a guarantee
A stop-loss becomes a market order once the trigger price trades. It then fills at the next available price. In ordinary conditions those are nearly the same. Over a weekend gap, a central bank surprise or a thin-liquidity hour, they are not, and the difference is entirely at your expense.
The practical consequence: the one percent you calculated is a best case, not a maximum. Any position held over a weekend or through a scheduled release should be sized on the assumption that the stop fills worse than requested — sometimes materially worse.
Spread and slippage are part of the risk, not a footnote
Every trade starts at a loss equal to the spread, and spreads widen exactly when volatility rises, which is exactly when strategies tend to trade. A plan calculated on typical spreads is calculated on the conditions where it matters least. Both costs are set out in more detail under spread, slippage and liquidity.
Leverage does not change your risk — until it does
Leverage is frequently blamed for losses it did not cause. If position size is set by the stop distance and the amount risked, leverage is irrelevant to the loss on any single trade; it only determines how much margin that position occupies.
Where leverage genuinely bites is margin. High leverage lets you hold positions whose combined margin requirement leaves little free equity. A drawdown then triggers a margin call that closes positions at the worst possible moment — not because the strategy was wrong, but because there was no room to be temporarily wrong.
Correlation makes one bet look like several
Risking one percent on each of five positions is only risking five percent in total if those positions are independent. EUR/USD, GBP/USD and AUD/USD long are substantially the same trade: short the dollar. A dollar rally takes all three together.
The workable habit is to think in terms of exposure to a driver — the dollar, risk sentiment, a single central bank — rather than counting open tickets.
What automation changes, and what it does not
An Expert Advisor enforces the parts of a risk plan that human discipline reliably fails at: it sizes every position by the same formula, places the stop before the entry is live, and does not widen a stop because a position is “about to turn around”. That is a genuine advantage, and it is the least glamorous part of automation.
What it does not change is everything above. An automated system faces the same gaps, the same spread widening and the same correlation. It will execute a flawed risk model with perfect consistency. Automation removes the emotional failure mode and leaves every structural one intact.
Testing whether a plan actually holds
Three questions worth answering before risking capital, all of which can be answered on paper:
- What does a realistic losing streak do? Not the worst you have seen — the worst the strategy’s win rate makes plausible over a few hundred trades. If that number ends the account, the position size is wrong regardless of the expected return.
- What happens if every stop fills a little worse? Re-run the arithmetic with each loss slightly larger and each win slightly smaller. A plan that only works on ideal fills is not a plan.
- What is the largest simultaneous exposure? Take the maximum number of positions the rules permit, assume they are correlated, and check whether that combined loss is acceptable. It usually is not, which is why a cap on concurrent positions matters more than the per-trade percentage.
The Bank for International Settlements triennial survey is a useful reference for which pairs carry genuine depth, since liquidity is what determines how far the gap between plan and execution can open.
The short version
Risk management is usually taught as a set of percentages, which makes it sound solved. It is better understood as a set of assumptions — that stops fill where you put them, that spreads stay near their average, that separate positions are separate bets — each of which fails under exactly the conditions that produce large losses. The rules are the easy half. Knowing where they stop being true is the half that protects the account.