ATR Position Sizing: Contract Value, Trading Costs and Gap Risk

A stop placed two ATRs away does not, by itself, define how much an account can lose. ATR supplies a distance in price units. Position quantity, contract value, currency conversion and execution costs turn that distance into money.

The useful question is therefore not simply which multiplier to choose. It is whether the exit rule and the position calculation describe the same risk. Separating them also makes a losing trade easier to diagnose: the market thesis may have failed, or a sensible stop may have been paired with an oversized position.

What ATR measures, including gaps

True range, or TR, is the largest of the current high minus low, the absolute difference between the high and the previous close, and the absolute difference between the low and the previous close. With a previous close of 100, a high of 106 and a low of 103, those values are 3, 6 and 3. TR is 6. Using the intrabar range alone would omit the displacement from the previous close.

A common 14-period implementation starts with the arithmetic mean of 14 valid true ranges, then applies Wilder smoothing: new ATR = (previous ATR × 13 + current TR) ÷ 14. If ATR was 2 and the next TR is 6, ATR becomes approximately 2.2857. A subsequent TR of 2 leaves it around 2.2653.

That lag works in both directions. Before a shock, historical ATR may understate the move about to occur; after conditions settle, it can remain elevated. ATR also has no directional sign: a rally, a sell-off or a wide bar that closes unchanged can all increase it. Fourteen five-minute bars and fourteen daily bars measure different horizons. Match the timeframe, session, price source, smoothing and initialisation before comparing readings.

Define the exit before sizing the trade

An ATR multiple can be a tested volatility exit, or help assess the buffer around a structural invalidation level. These are different rules. Suppose a proposed long entry is 100, the lower edge of support is 96.5 and two ATRs equal 2. A stop at 98 may trigger while the support structure remains intact. That may suit a tighter volatility rule, but it is not an exit only after support fails.

If the trade depends on support, specify what would invalidate that structure and whether it needs a buffer. If it uses a mechanical ATR exit, test its suitability for the market and holding period. The guide to support, resistance and false breakouts provides the structural context. Moving the stop closer merely to afford a larger position reverses the logic.

The basic calculation is: risk budget = account equity × chosen risk fraction; planned risk per tradable unit = stop distance × monetary value of one price unit + additional costs; maximum quantity = budget ÷ unit risk, rounded down to the permitted increment.

The 0.5% fraction below is an illustrative assumption, not a universal allocation. Nor does it guarantee a 0.5% loss ceiling. Minimum charges, fee tiers and nonlinear market impact require a fresh total-cost calculation for the candidate position, rather than a fixed cost per unit.

A stock example: wider stops require fewer shares

Consider a hypothetical account with 100,000 currency units of equity and a 500-unit budget. Entry is 50 per share; ATR from completed daily bars is 1.2. A two-ATR exit is 2.4 away. Allowing another 0.10 per share for round-trip charges and adverse slippage gives 2.50 of planned risk per share. The position is 500 ÷ 2.50 = 200 shares, worth 10,000 at entry.

This initially assumes whole-share dealing. If the relevant venue instead requires purchases in increments of 100 shares, 200 still fits. A calculated 240 shares would need rounding down to 200. These increments are hypothetical, not a statement of any market's current rules.

Now let ATR double to 2.4. The exit distance becomes 4.8 and unit risk, including the same cost allowance, becomes 4.90. The budget permits about 102 shares, or 100 under the assumed increment. Keeping 200 shares while widening the stop raises planned risk to 980. A rising ATR does not authorise an increase in the loss budget, including after a position has already been opened.

Futures: points are not ticks

Take a fictional contract worth 10 currency units per one-point move. Its minimum tick is 0.2 points, so each tick is worth 2. With ATR of 3 points, a two-ATR stop is 6 points or 30 ticks away. Price risk is 60 per contract. Add a hypothetical 10 for round-trip charges and slippage, bringing unit risk to 70.

A budget of 1,000 permits floor(1,000 ÷ 70) = 14 contracts, for planned risk of 980. Mistaking six points for six ticks would produce an erroneous price-risk figure of 12. Checking contract units can matter far more than changing the ATR lookback from 14 to 20.

Fourteen is only the ceiling imposed by this particular budget. Margin requirements, available-cash reserves, market depth and other position limits may require less. Posting margin does not eliminate price risk or guarantee that losses cannot exceed the initial margin.

FX and gold: verify the exposure behind the quote

For EUR/USD in a dollar account, a position of EUR 100,000 gains or loses USD 10 for a 0.0001 move. Assume ATR is 0.0008, or eight pips. A two-ATR distance is 16 pips, giving USD 160 of price risk. With USD 10 of assumed round-trip costs and slippage, the total is USD 170.

A USD 200 budget implies 200 ÷ 170 × 100,000, or about EUR 117,647. Assuming EUR 1,000 dealing increments and costs proportional to quantity, round down to EUR 117,000. Planned risk is USD 198.90. Writing the base-currency amount avoids assuming that every provider uses the same definition of a lot.

If account and quote currencies differ, convert the resulting profit or loss; some pip values also vary with exchange rates. Gold poses a similar unit problem: a USD 1 move per ounce has ten times the effect on 100 ounces as on ten. Spot, futures and CFDs can show similar prices while having different contract sizes, minimum quantities and charges.

For a relative-volatility comparison, ATR ÷ closing price × 100% gives 2% for ATR of 2 at a price of 100, but 1% for ATR of 20 at 2,000. The larger nominal range is not necessarily the larger proportional move. Account risk still depends on quantity and monetary exposure.

Why a 500-unit budget can become a 700-unit loss

Return to the 200 shares bought at 50 with a stop trigger of 47.6. If overnight news leaves the first available execution price at 46.5, the price loss is (50 − 46.5) × 200 = 700, before fees. A stop-market trigger is not a guaranteed fill. A stop-limit order can restrict the acceptable price, but may leave the position unfilled.

The 500 figure is planned risk under execution assumptions, not a maximum under every scenario. Positions spanning closures, company announcements or major economic releases need separate consideration of gaps, thin liquidity and trading interruptions. If a plausible exit range cannot be estimated reliably, reducing exposure or declining the position is part of risk budgeting.

Portfolio concentration adds another layer. Three positions each budgeted at 500 may all depend on a weaker dollar or rising equities. Their individual calculations can be correct while their common exposure produces simultaneous losses. Historical correlations do not guarantee diversification under stress.

Audit the source of the loss

Record the ATR available from completed bars before entry, the exit rule, contract value, cost assumptions, rounded quantity and the difference between planned and realised loss. A signal generated at the close should be evaluated using the next tradable price; the signal's closing price is not automatically available for execution.

Frequent exits during ordinary fluctuations suggest a mismatch between the exit rule and horizon. Persistent budget overruns point toward units, sizing or execution assumptions. Losses clustered on the same day point toward shared exposures. ATR makes those distinctions measurable; no multiplier can turn them into a guaranteed boundary.