Position Sizing: Put the Loss Boundary Before the Opportunity

Educational use only: the examples below are simplified demonstrations, not financial advice or a recommendation to trade. Real losses can exceed a planned amount because of gaps, slippage, fees, leverage and currency conversion.

Position sizing answers a more useful question than “How confident am I?” It asks how much exposure is consistent with a predefined loss boundary if a scenario is wrong. Good sizing starts with the portfolio and works down to the instrument. It does not start with the largest quantity a platform allows.

Four inputs, in the correct order

1. Account or portfolio reference value: the amount used for the risk calculation.
2. Loss budget: the maximum planned loss for this single scenario.
3. Entry assumption: the price used for planning, not a guaranteed execution price.
4. Invalidation point: the price at which the original scenario is considered wrong.

The invalidation point should come from the logic of the scenario before quantity is considered. Moving it merely to create a larger position reverses the process. A stop order may help implement a boundary, but it cannot guarantee a particular fill.

The basic calculation

For a simple unleveraged illustration, planned loss per unit is the absolute difference between the assumed entry and invalidation price. The loss budget is the reference value multiplied by the chosen risk percentage. The theoretical number of units is the loss budget divided by planned loss per unit.

Suppose a fictional learner uses a reference value of 10,000 units of account currency, a 0.5% learning-case loss budget, an assumed entry of 50 and an invalidation point of 48. The loss budget is 50, the planned loss per unit is 2 and the frictionless result is 25 units. Those numbers are deliberately neutral and do not imply that 0.5% is suitable for anyone. The result is a starting estimate, not an executable instruction.

Add the frictions the formula ignores

• Spread and commission: entry and exit costs reduce the amount available for price risk.
• Slippage and gaps: the actual exit may be worse than the invalidation price.
• Contract specifications: futures, options, CFDs and leveraged products can have multipliers and non-linear behavior.
• Currency conversion: the instrument and account may use different currencies.
• Minimum increments: a platform may require whole shares, lots or contract sizes.
• Liquidity: displayed prices may not be available for the intended quantity.

A conservative worksheet can reserve part of the loss budget for costs and adverse execution, then round quantity down to a permitted increment. If the instrument is complex, the basic unit formula is not enough.

Think at portfolio level

Five individually small positions can represent one large risk if they respond to the same driver. For example, several assets may all depend on the same currency, interest-rate expectation, sector or broad risk sentiment. Before adding exposure, group existing positions by underlying driver and imagine them moving together. Diversification by ticker name is not necessarily diversification by risk.

Also distinguish open risk from realized loss. An account can carry several scenarios whose combined invalidation losses are larger than intended. A portfolio risk budget should consider simultaneous adverse outcomes, not only one position at a time. The FINRA overview of asset allocation and diversification provides useful context for examining combined exposure.

A repeatable worksheet

1. Write the scenario and the evidence that would invalidate it.
2. Record the reference value and the maximum planned loss in both percentage and currency terms.
3. Calculate theoretical unit risk and quantity.
4. Estimate fees, spread, slippage, conversion and contract effects.
5. Round down and check the new planned loss.
6. Review correlation with existing exposure and any event risk during the holding window.
7. Record the result before action; do not reconstruct the plan afterward.

Common sizing errors

Frequent mistakes include treating a close stop as proof of low risk, increasing size after losses to recover quickly, using the same quantity for instruments with different volatility, and ignoring the possibility of an unavailable exit. Confidence is not a mathematical input. Neither a long research process nor a strong opinion reduces the amount an instrument can move.

The Investor.gov guide to risk is a useful companion when identifying risks that a single calculation cannot capture. The disciplined conclusion is simple: define how wrong can be handled before calculating how much right might earn. Recalculate whenever the entry assumption, invalidation point, portfolio value or related exposure changes; an old quantity does not belong to a new scenario.