Forex Risk Management: The Rules and the Maths Behind Them

Forex risk management is the set of rules that decides how much you can lose, on one trade and in total, before you place any trade. Its core is simple: risk a small, fixed share of the account on each trade, put a stop-loss on every position, and calculate the lot size from the distance to that stop.

• Risk per trade, commonly 0.5–2% of the account. See risk per trade.
• A stop-loss on every trade, placed where the trade idea is wrong rather than where it suits the size. See stop-losses.
• A lot size calculated from the stop, using the lot size formula.
• A reward:risk ratio that fits your win rate. See the risk-reward ratio guide.
• Limits on total open risk and on daily or weekly losses, with correlated positions counted together. See correlation.

This is what 1% risk means in lots on a $10,000 account in US dollars:

The same rule applies to small accounts, where the minimum lot size gets in the way. With $500, 1% is $5, and a $5.00 stop on gold allows exactly 0.01 lots (1 oz, worth $1 per $1.00 move). A wider stop at that balance means risking more than 1% or skipping the trade.

Risk per trade is the amount you lose if a trade reaches its stop-loss, expressed as a share of your account. It is the single setting that decides how quickly a run of bad trades can damage an account, and unlike win rate, market conditions or execution, it is entirely under your control.

The "1% rule" you will see everywhere is a convention, not a law. Most published guidance for discretionary traders falls between 0.5% and 2% per trade. The right figure depends on how many trades you take, how correlated they are, how long your losing streaks get, and how deep a drawdown you can sit through without changing your behaviour. What matters more than the exact number is that it is decided before the trade and turned into a lot size with the lot size formula.

At 1% risk, twenty consecutive losses, an extreme but not impossible run, leave 82% of the account. At 5% they leave 36%, and at 10%, 12%. The difference between those rows is not skill; it is a single input.

R is the amount you risk on a trade. A trade that loses its full stop is −1R; one that makes twice the risk is +2R. Measuring results in R rather than in pips or money makes trades of different sizes, instruments and account balances comparable, and it separates the quality of a strategy from the size you trade it at.

A strategy that wins 45% of the time with an average win of 2R and full-stop losses has an expectancy of 0.45 × 2 − 0.55 × 1 = +0.35R per trade. At 1% risk that is an average of about 0.35% of the account per trade before costs, or roughly 35R over 100 trades.

Expectancy says nothing about the path. The same +0.35R strategy will go through long losing streaks, and the amount you risk per R decides whether you survive them. That is what the rest of this guide is about.

where R is the reward:risk ratio: the distance to the target divided by the distance to the stop. The higher the ratio, the fewer trades need to win:

A higher ratio is not automatically better. Distant targets are reached less often, so the win rate falls as R rises. What matters is whether your actual win rate, at the targets you actually use, sits comfortably above the break-even rate. The risk-reward ratio guide works through examples on EUR/USD, USD/JPY and gold.

Costs move the break-even point

Spread and commission are paid on every trade, winning or losing, so they shrink the reward and enlarge the risk. Take a 20-pip stop and a 40-pip target (1:2) on EUR/USD with 1.9 pips of all-in costs: a 1.2-pip spread plus $7 round-turn commission. The effective risk becomes 21.9 pips and the effective reward 38.1 pips, a ratio of 1.74, and the break-even win rate rises from 33.3% to 36.5%. Halve the stop and target to 10 and 20 pips and it rises to 39.7%. Short-term strategies need a noticeably higher win rate for the same nominal ratio.

A drawdown is the fall from an account's peak to its subsequent low. Because the recovery is earned on a smaller balance, getting back to the peak always takes a larger percentage gain than the loss:

Up to about 20% the asymmetry is mild. Beyond 30–40% it becomes punishing: a 50% loss needs the account to double just to get back to where it started. This is the strongest argument for modest risk per trade. It keeps ordinary losing streaks in the part of the table where recovery is realistic.

Drawdowns also take time. A strategy averaging +0.35R per trade at 1% risk, about 0.35% per trade, needs around 30 trades on average to recover a 10% drawdown, and more if you cut risk while recovering, as many traders do.

People consistently underestimate how long losing streaks get over a long series of trades. The table shows the probability of at least one losing streak of a given length somewhere within 100 trades, assuming each trade is independent.

A strategy with a 40% win rate, perfectly viable at 1:2 or better, has roughly even odds of eight losses in a row within 100 trades and about a one-in-five chance of ten. At 2% risk a ten-trade streak is an 18% drawdown; at 1%, under 10%. Set your risk so that the streak you should expect is survivable, financially and psychologically, and so that you can tell the difference between a normal streak and a strategy that has stopped working.

Risk of ruin is the probability that an account falls to a level you define as "ruined", often a 50% drawdown rather than zero, before it grows. It depends on three things: the strategy's edge (win rate and reward:risk), the risk per trade, and where you draw the ruin line.

For a fixed amount risked per trade, the classic gambler's-ruin result gives the probability as z^C, where C is the number of losing trades between you and the ruin level (ruin drawdown ÷ risk per trade) and z is a number between 0 and 1 set by the edge. You don't need the algebra to use the intuition: the probability falls exponentially as C grows, so halving the risk per trade squares an already small probability.

The same strategy is almost bullet-proof at 1% and has better than a one-in-three chance of halving the account at 10%. With no edge, meaning zero or negative expectancy, the risk of ruin is 100% at any size given enough trades; smaller risk only slows the decline.

Two caveats. The model assumes your win rate and reward:risk are known exactly, when in practice they are estimates from a limited sample that may well be optimistic. And it assumes each trade is independent with a fixed risk; correlated positions and slippage make reality worse. Treat the figures as a lower bound.

where W is the win rate and R is the average win divided by the average loss. f* is the fraction of capital to risk per trade that maximises long-run growth, if and only if W and R are known exactly and never change.

Full Kelly figures look absurd next to the 1% convention, and for trading they are, for three reasons. First, growth at full Kelly comes with very deep drawdowns: in the standard continuous-time approximation, the chance of the account ever falling to half its starting value is 50%. Second, overestimating your edge, which is easy with a few hundred trades, pushes you past the Kelly fraction, where growth falls and the risk of ruin climbs sharply. Third, the formula ignores costs, correlation and the way people trade differently in a deep drawdown.

That is why practitioners who use Kelly tend to use a quarter or half of it, and why it is better read as a ceiling than a target: if your risk per trade is above a quarter of your estimated Kelly fraction, you are probably taking more risk than your edge justifies. A Kelly fraction of zero or below means the strategy has no edge to size.

Risk per trade only works if trades are independent, and often they aren't. Long EUR/USD, long GBP/USD and short USD/CHF are three positions but largely one view: that the US dollar will fall. If each risks 1%, a broad dollar rally can hit all three stops together, a 3% loss on what was effectively one trade.

Correlation (ρ) measures how closely two instruments move together, from +1 (in lockstep) to −1 (mirror images). For two positions risking the same amount, the combined volatility relative to a single position is √(2 + 2ρ). With ρ = 0.8 that is 1.90, barely less than the 2.0 of two perfectly correlated positions. Diversification between highly correlated currency pairs is mostly an illusion.

Common ways to deal with it:

• Set a cap on total open risk across all positions (sometimes called portfolio heat), for example 3–5% of the account, and count correlated positions together.
• When two positions express the same view, split your normal risk between them rather than doubling it.
• Remember that correlations are unstable. Instruments that usually move independently can move together in a sell-off, often exactly when it matters most.

The order of operations matters. Decide where the trade is wrong (beyond a swing point, outside the normal noise measured by ATR, or wherever your strategy says) and only then size the position so that distance costs your chosen risk. Choosing a size first and squeezing the stop to fit puts the stop where the market routinely reaches it.

• Don't widen a stop on an open trade without reducing the size. Moving a stop from 20 to 40 pips at the same size doubles the risk.
• Know what your stop actually guarantees. An ordinary stop becomes a market order when triggered; in fast markets and over weekends it can fill well beyond its price. Guaranteed stops, where offered, cost a premium.
• For a long position, the stop is triggered by the bid while your entry was at the ask, so the spread is part of the loss at the stop.

Funded-account (prop firm) challenges turn risk management into hard rules: break one and the account is closed, whatever your long-run edge. Two limits are almost universal:

• Daily loss limit, commonly 4–5% of the starting account size, measured from the balance (or equity, or the higher of the two) at the start of each day. It is breached the moment equity, including open losses, touches the level.
• Maximum loss, commonly around 10% of the account size. It may be static, a fixed floor, or trailing: the floor rises with the highest balance or equity reached, sometimes only at the end of each day, and often stops rising once it reaches the starting balance.

Worked example

A $100,000 account with a daily limit of 5% of the starting account size, measured from each day's opening balance, and a 10% maximum loss. Today opens with a balance of $102,000.

The tightest limit binds: here, the daily limit. Note how the trailing version works: profit that has been made and then given back still raised the floor, so the trailing account has $8,000 of room where the static one has $12,000.

Sizing for the limits

Divide the room you have by your risk per trade to see how many full losses fit. At 1% ($1,000) per trade, five losses breach a 5% daily limit, and fewer if open trades are in drawdown at the same time or a stop slips. Many funded traders therefore risk 0.25–0.5% per trade, cap the number of trades or the total open risk per day, and stop for the day after a set loss well inside the limit. Size from the room left, not from the account size: with $3,000 of daily room remaining, a $1,000 trade is a third of what's left.

Some firms also apply a consistency rule, which caps how much of the total profit any single day may contribute. That is a reason to keep size steady rather than pressing after a good day.

A risk plan is short. It answers in advance the questions that are hard to answer well in the middle of a losing streak:

1. Risk per trade, as a percentage or a fixed amount, and how it is turned into lots.
2. Maximum total open risk, with correlated positions counted together.
3. A daily or weekly loss at which you stop trading for the rest of that period.
4. A drawdown at which you cut risk per trade (for example, halving it after a 10% drawdown) and a deeper level at which you stop and review the strategy.
5. Rules for stops: where they go, and that they are never widened without resizing.
6. Which events (central-bank decisions, major data releases, weekends) you will not hold positions through, or will hold only at reduced size.

The numbers in the plan are yours to choose. The point of the maths above is to make sure you choose them knowing what they imply.