Position sizing: start with the loss

Three numbers, three different meanings

A position has a purchase value, a required cash commitment, and an exposure to loss. These are not interchangeable. Buying 100 shares at $40 creates $4,000 of market exposure. In a fully paid cash purchase, $4,000 is committed before fees. A leveraged product might require a much smaller deposit for similar exposure, but a small deposit does not make the underlying price movement smaller.

The useful starting question is not how much a platform allows you to buy. It is what a specified adverse scenario would do to the account. A scenario can be a planned exit, a sharp overnight gap, or a failure of the counterparty. Different scenarios capture different risks; no single sizing formula describes all of them.

Build the arithmetic before opening an order

For a simple long cash-share example, planned loss per share equals entry price minus assumed exit price. Divide a chosen hypothetical loss budget by this distance to obtain an initial share count. Add estimated commissions, spread and slippage, then round down to the permitted trading increment. Futures, currency contracts and derivatives need their own contract multiplier and account-currency conversion.

Imagine a simulated $10,000 account with a $100 planning budget. Entry is $25 and the intended exit is $24.20, a distance of $0.80. Reserving $10 for estimated round-trip costs leaves $90 for the price movement. That gives 112 whole shares when rounded down: $89.60 of price loss plus the $10 allowance. The $100 figure is an arithmetic assumption, not a recommended risk percentage.

Costs do not always behave like a single fixed allowance. Some are charged once per order, some per share or contract, and some as a percentage of notional value. If the cost grows with quantity, subtracting a fixed estimate and then increasing size can understate the loss. Write the cost schedule into the calculation before rounding. When the estimated fixed costs alone use up the planning budget, the arithmetic supports no position under those assumptions.

Stress the assumption that matters most

The calculation works only if the assumed exit is attainable and the cost estimate is adequate. Suppose an announcement sends the next available price to $22.80. The same 112 shares lose $246.40 on price alone. A stop order does not turn $24.20 into a contractual guarantee. Very illiquid markets can make the gap between a model and execution particularly large.

Run another scenario with a wider spread, an exchange outage, or a currency movement affecting the value of the loss. With leverage, include the possibility of forced liquidation before your intended exit. Product terms and local protections determine whether a deficit can remain; do not infer protection from a platform interface or marketing label.

Count the positions already open

Five separate trades each carrying a $100 planned loss are not a $100-risk portfolio. Nor are they independent simply because their ticker symbols differ. A long technology share, a technology index fund and a related leveraged contract may all respond to the same event. Write down aggregate planned loss and a common-shock scenario before looking at a new trade in isolation.

Position sizing is a constraint, not evidence that a strategy works. Smaller positions can slow losses from a poor process but cannot manufacture a profitable edge. A paper exercise is complete when its units, costs, assumptions and failure scenarios are explicit enough that someone else can reproduce the calculation.

A useful final check is dimensional: shares multiplied by dollars per share gives dollars; contracts multiplied by ticks and dollars per tick gives dollars. If the answer is in a foreign currency, convert it using an explicit assumed rate. Writing these units beside every number catches errors that a polished calculator interface can conceal, especially when switching between instruments with similar-looking prices but different multipliers.

Let every constraint have a say in the final size

Consider a separate hypothetical share trade with a $100 planning budget, an $0.80 adverse price distance, $0.07 of estimated round-trip costs per share and $8 of fixed round-trip costs. The loss model is quantity × ($0.80 + $0.07) + $8. Dividing the remaining $92 by $0.87 gives about 105.75 shares, so the largest whole-share quantity within that model is 105. Its planned loss is $99.35; rounding to 106 raises the modeled loss to $100.22.

That answer is only a risk-based ceiling. A fully paid purchase also needs enough available cash to cover its price and entry charges. An exposure limit, existing related positions, a minimum order increment or limited available liquidity can impose a smaller ceiling. The feasible quantity must satisfy all applicable constraints, not whichever one produces the largest number. A zero-size result is a meaningful finding, not a calculation that needs to be overridden.

Changing the assumed exit changes size dramatically. With the same $100 budget, $8 fixed cost and $0.07 per-share cost, doubling the adverse distance from $0.80 to $1.60 reduces the whole-share ceiling to 55. Moving the exit closer merely to justify more shares reverses the logic of the exercise. The exit assumption should describe the hypothesis and plausible execution; size then follows from the loss model, subject to the other constraints.

Before recording the final quantity, write a second loss scenario that ignores the intended exit and uses an adverse executable price. Keep both numbers visible. The planned-loss calculation answers what happens if the modeled process works as assumed; the stress calculation answers what happens when a crucial assumption fails. Neither should be relabeled as a guaranteed maximum unless the product itself creates an enforceable payoff limit and its terms genuinely support that claim.