Position Sizing With Variable Stops: Keeping Risk Constant

[FXS Prop Desk]

Many scalpers use structure-based stops, so the stop distance changes every trade. Keeping money risk constant requires recalculating size each time.

Formula:

Code: Select all

lots = (account × risk%) / (stop_pips × pip_value_per_lot)

Example: 20,000 account, 0.5% risk = 100. Stop 6 pips on EURUSD (pip value ≈ 10 per lot) → 100 / 60 = 1.66 lots. Stop 12 pips → 0.83 lots.

A simple way to measure the effect is to tag each trade with the number of minutes since the previous loss. If trades taken within five minutes of a loss perform clearly worse, the pause rule is supported by your own data.

Things to watch:

• Pip value changes with the quote currency (e.g. GBPJPY, AUDCAD).
• Add expected spread and commission to the stop distance, or the real risk is higher than planned.
• Round lots down, not up.

Do you calculate size manually, with a script, or with a platform tool?

[Fairman]

Add expected spread and commission to the stop distance, or the real risk is higher than planned.

That point matters more the tighter the stop is, so it's worth putting numbers on it. Using your example, a 6 pip stop with 0.2 spread, 0.7 commission and say 0.3 average slippage on stops is really 7.2 pips of risk. 100 / 72 gives about 1.39 lots instead of 1.66, so the uncorrected size would risk roughly 0.6% instead of 0.5%. On the 12 pip stop the same costs only push it to 13.2 pips, a much smaller distortion. So scalpers with the tightest stops are the ones most likely to be quietly over-risking, which is a decent argument for building the cost into whatever calculator or script does the sizing.

[Shadow Trader]

The weak spot I hit with this formula is the very tight stop. A 2 pip stop on the same 20,000 account gives 5 lots, and with Fairman's cost figures the real distance is 3.2 pips, so the uncorrected size risks about 0.8 percent instead of 0.5.

So my script has two guards besides cost. It refuses stops under 4 pips on EURUSD and caps size at 2 lots.

Lot step is the opposite problem on small accounts. On 2,000 at 0.5 percent with an 18 pip stop the answer is 0.055 lots, rounding down gives 0.05, and the trade risks 9 dollars instead of 10, which I can live with.

For JPY pairs I read the tick value from the symbol specification at the moment of the order. With USDJPY at 150 a GBPJPY pip is about 6.67 dollars per lot, so a fixed 10 in a spreadsheet would undersize every trade by a third.

I think the paragraph about minutes since the previous loss belongs to a different thread. It is a good test, just unrelated to sizing.

Do you size off balance or equity when a trade is already open? I use balance minus open risk, so two trades together never exceed the plan.