Risk management basics — the foundations every forex trader needs

A profitable year after an earlier loss does not isolate the effect of a new risk rule. Review the recorded initial risk, realised R-multiples, stop distance, costs and combined exposure. This article uses worked models to show how those quantities differ and how to document them.

Part 1 of 6What a risk plan should specify

ESMA’s 2018 retail-CFD intervention cited historical analyses in which 74–89% of retail accounts lost money. Current provider-specific warnings vary. These historical figures do not isolate sizing, analysis, costs or another factor as the cause of any particular account loss.

The standard EU retail forex-CFD intervention model limits leverage to 30:1 for major pairs and 20:1 for other currency pairs, subject to applicable national measures and client status. Poland’s qualified experienced-retail category permits different initial margin on specified instruments. Identify the contracting entity, covered account and category before applying a margin or negative-balance rule.

Part 2 of 6The one-percent rule — what to calculate and how

A 1% rule defines initial planned risk, not a guaranteed maximum execution loss. In a hypothetical model where each actual loss equals the selected percentage of then-current equity, with no additional costs, ten 1% losses leave 0.99¹⁰ ≈ 90.44%; ten 2% losses lose about 18.29%; ten 5% losses lose about 40.13%. Recovering an exactly 40% drawdown needs 1/0.60 − 1 ≈ 66.67% on the remainder.

Use the recorded sizing rule consistently enough to explain each decision, and preserve changes in the journal. Consistency alone does not make a 1.5% fraction suitable or prove superiority over variable sizing: the underlying outcome distribution, execution and total exposure matter. The guide on the pip and its value across pairs explains units for the calculation.

Part 3 of 6R-multiple and the risk-to-reward ratio

An R-multiple expresses net realised P/L relative to the initial risk reference. With initial risk USD 100, net profit USD 250 is +2.5R. A loss is−1R only if its actual net amount is100USD; slippage, gaps and costs can make it worse than−1R.

Use stated probabilities, average realised payoffs and costs together. In a hypothetical binary model with wins +2R and losses −1R, a 40% win probability gives 0.40×2−0.60×1=+0.20R before costs; a 1:1 model breaks even at 50% before costs. No target ratio alone guarantees profitability or establishes which ratios most successful retail strategies use.

Part 4 of 6Position sizing — the formula

For a stated linear price-loss model, volume equals the account-currency budget divided by the product of stop distance and pip value for one specified volume unit. Capital and the chosen fraction can be recorded before entry; the actual fill, exit and conversion may differ later. Verify the contract specification, volume step and costs rather than assuming every input remains fixed.

Hypothetical USD 10,000 base, 1% price-loss budget, 20-pip entry-to-exit distance and USD 10 per pip per standard lot give 100/(20×10)=0.50 standard lots. If mini and micro units are defined as 0.1 and 0.01 standard lots, that is five minis or fifty micros. One standard lot models USD 200 loss instead of 100 under these same fixed assumptions; no realised fill or psychological response is guaranteed.

Part 5 of 6Stop loss — where to place it

Stop distance should be evaluated in the context of the instrument, timeframe and entry rule. ATR is a historical range measure, not a prediction or a fixed currency-pair ratio. This article supplies no dated data establishing that GBP/JPY volatility is twice EUR/USD volatility.

An ATR multiple is a testable stop-placement rule, not a universally suitable buffer. State the timeframe, period setting, multiplier and invalidation convention before evaluating it. This article does not establish fixed12–25pip or30–60pip ranges for future sessions. See ATR as a volatility measure.

Write how exit levels may change before placing the order. Moving a long entry 1.0850 stop from 1.0830 to 1.0820 increases its planned price-loss distance from 20 to 30 pips at unchanged size, a 50% increase before costs. A trailing change also alters the exit rule and needs evaluation; it is not automatically sensible for every strategy. Do not move a stop merely to hide a budget breach.

An ordinary stop is an intended exit, not a guaranteed fill price or maximum loss. The calculated 1% is initial planned risk, not a statistical expected value. For a covered retail CFD account, negative-balance protection limits aggregate liability to dedicated CFD funds; it does not preserve the deposit or apply to every contract.

Part 6 of 6Drawdown and portfolio correlation

For a stated equity series and window without cash flows, drawdown is the decline from the running peak to a later observation. A hypothetical path 10,000→12,000→10,800 gives (12,000−10,800)/12,000=10%. Maximum drawdown is the largest such decline in the chosen window. An exact 30% decline needs about 42.86% recovery and 50% needs 100%; these identities do not set universal 15–20% tolerance or a 25% pause threshold.

Positions such as long EUR/USD, GBP/USD andAUD/USD share a USD exposure, but their return correlations vary with dates, frequency and conditions. This article provides no dataset establishing a universal 0.80–0.90 correlation. Three 1% planned risks sum to 3% regardless of independence; correlation affects joint outcomes, and actual losses may exceed planned risks. Any portfolio limit is a chosen rule, not a guaranteed maximum.

A deeper technical treatment of position sizing, R-multiple and the mathematics of ruin sits in the risk management section in our library, including templates for sizing positions on multi-currency accounts.

What to do tomorrow

1. State the sizing unit. Divide the account-currency initial risk budget by stop distance in pips multiplied by pip value for one specified volume unit. If pip value is per standard lot, the quotient is standard lots; if it is per micro-lot, the quotient is a number of micro-lots. Round to the broker’s volume step and check actual execution and costs.
2. Review realised R-multiples. For each trade record initial monetary risk and actual net P/L in the same currency; R-multiple is net P/L divided by that initial risk. Entry price minus stop price is a price distance, not USD risk without volume, pip value and conversion. Twenty observations provide a sample estimate, not a universal +0.3R quality cutoff.
3. Inspect currency legs and combined scenarios. Long EUR/USD, GBP/USD and AUD/USD share short-USD exposure but are not one identical position. List amounts, directions and possible joint losses. A rule limiting exposure must specify its basis and actions; neither two positions nor 3% is a universal safe limit. The article on hedging the positions explains why adding a hedge can introduce costs and other risks.
4. Define daily and monthly loss procedures. Any example limits must specify their capital base, treatment of realised losses and remaining exposure, and actions for existing positions and orders. Closing the platform alone leaves exposure active; a threshold does not guarantee loss stops exactly there.
5. Record R-multiples with their definitions. Use date, instrument, direction, initial monetary risk, net P/L and realised R, all in a stated account currency. Store execution deviations and rule changes too. Fifty observations can describe that sample, but do not automatically establish future expectancy or validate a setup.