Position sizing: the decision that precedes every trade

Most discussion of trading concerns what to buy. Far less concerns how much. Yet the size of a position determines how much a wrong call costs you, and wrong calls are guaranteed. Sizing is the one decision where a rule can be written in advance and followed without judgement.

Start from the loss, not the gain

A position size follows from two numbers: the amount of capital you are prepared to lose if the idea fails, and the price at which you will concede that it has failed. Divide the first by the distance to the second and you have the number of shares. Nothing about the expected upside enters the calculation.

Why fixed-fraction risk works

Risking a fixed fraction of capital per idea has two useful properties. Losses shrink in dollar terms as capital shrinks, which makes ruin mathematically unreachable. Wins compound in dollar terms as capital grows, without any change to the rule.

• The fraction is small. Practitioners commonly cluster between 0.25% and 1.0% per idea. Above 2% the sequence of losses that ends a strategy becomes uncomfortably short.
• The exit level is set before entry and is not moved further away afterwards. Moving it closer is permitted.
• Correlated positions share a budget. Three semiconductor names sized independently are one large semiconductor position.

What sizing cannot do

Sizing does not improve the quality of an idea. A portfolio of poorly chosen positions, sized correctly, loses money slowly rather than quickly. That is still a useful property: it buys time to discover that the process is not working.

A note on volatility-scaled sizing

Some practitioners replace the exit distance with a volatility estimate, such as an average true range, so that positions in calmer names are larger. The logic is the same.