The Central Role of Position Sizing in a Trading System
The study notes make a direct claim: position sizing is the most important part of any trading system, because it is only through position sizing that a trader can reach an objective. Expectancy, in this framing, is a rough estimate of a system's true potential. Once a suitable system has been developed, the notes say, the next decision is which calculation method will be used to determine position size. That decision is not a detail added at the end; it is the part of the system that converts a statistical edge into an actual equity curve. A simple way to see the point is to imagine two traders using identical entry and exit rules on the same market over the same period. One risks a small, fixed fraction of equity per trade; the other commits a large portion of the account to each signal. The signals are the same, but the outcomes can be completely different. The first trader may experience a moderate drawdown and continue trading. The second may be forced to stop after a normal losing streak, because the account no longer supports the minimum position. The edge did not change. The sizing did. This is why the notes treat position sizing as the goal-achievement mechanism. Entries and exits define when a trade is taken and when it is closed. Sizing defines how much of the account is exposed. A system with positive expectancy can still fail to meet an objective if the size is too small to compound meaningfully, or too large to survive the variance. The notes also warn that an inappropriate sizing method leads to ruin, however ruin is defined — whether that means losing half the capital or losing everything. The rest of this article examines the mechanics behind that warning, the conditions under which different sizing methods work, and the ways they are commonly misused.
Expectancy as a Rough Estimate of True Potential
Expectancy is usually defined as the average result per trade over a large sample. In its simplest form, it is the win rate multiplied by the average win, minus the loss rate multiplied by the average loss. A system with a 40% win rate, an average win of 2R, and an average loss of 1R has an expectancy of 0.2R per trade. Traders often express expectancy in R multiples, where 1R is the amount risked on a trade. This makes systems with different price levels and instruments comparable. The notes describe expectancy as a rough estimate of true potential. There are several reasons it is rough. First, it is estimated from a finite sample. A system that shows positive expectancy over 100 trades might have a wide confidence interval; the true value could be lower, zero, or negative. Second, market conditions change. A method that worked in a trending, high-volatility regime may behave differently in a range-bound, low-volatility regime. Third, real-world frictions — slippage, commissions, gaps, and missed fills — reduce paper expectancy. Fourth, the estimate depends on how trades are counted and how R is defined. Expectancy also says nothing about the sequence or the dispersion of outcomes. Two systems can share the same average trade result while having very different standard deviations. One may produce a smooth series of small gains and losses; the other may produce long losing streaks followed by large wins. Position sizing must account for the distribution, not just the mean. A sizing rule that is safe for a low-variance system may be fatal for a high-variance system with the same expectancy. Because expectancy is only an estimate, sizing rules need a margin for error. Assuming the estimate is precise is one of the most common errors in system design.
