Position sizing is the least glamorous part of trading and arguably the most important. Most traders spend a disproportionate amount of energy on finding entries and almost none on the question of how much to risk on each trade. The result is that even a profitable strategy can produce damaging drawdowns or inferior long-term returns because the sizing method is poorly suited to the strategy's characteristics.
There are several established money management frameworks used in professional trading. Understanding their properties - and their trade-offs - allows you to choose one that matches your strategy and your tolerance for variance in equity growth.
Fixed Fractional (Percentage Risk)
Fixed fractional sizing risks a fixed percentage of the current account balance on each trade. If you risk 1% per trade and your account is $10,000, you risk $100. If the account grows to $12,000, the next trade risks $120. If it falls to $8,000, the next trade risks $80.
This is the most widely recommended method for retail traders, and for good reason. The key properties are:
- Position sizes scale automatically with the account, both up and down
- The account cannot be theoretically wiped out by any sequence of losses, because each loss reduces the base that subsequent risk is calculated from
- Drawdown recovery is faster than with fixed lot sizing, because losses reduce position size, which slows further drawdown
- The compounding effect means profitable periods produce accelerating growth
The main limitation is that it does not account for the varying quality of setups. Every trade risks the same percentage regardless of whether the setup is marginal or high-conviction. For strategies with significant variance in setup quality, this is a meaningful constraint.
The Kelly Criterion
The Kelly Criterion is a formula from information theory, adapted for trading, that calculates the theoretically optimal fraction of capital to risk on each trade to maximise the geometric growth rate of the account. The formula is:
Kelly % = W - [(1 - W) / R]
Where W is the win rate and R is the average win-to-average loss ratio (reward-to-risk). A strategy with a 55% win rate and a 1.5:1 average reward-to-risk ratio produces a Kelly percentage of approximately 18.3% - meaning the formula suggests risking 18.3% per trade for optimal long-term growth.
In practice, full Kelly sizing produces equity curves with drawdowns that are psychologically and financially devastating for most traders. The theoretical optimum maximises the long-term growth rate but produces enormous variance along the way. Most quantitative traders who use Kelly apply a "fractional Kelly" approach - commonly half-Kelly or quarter-Kelly - which sacrifices some theoretical growth rate in exchange for dramatically reduced variance and drawdown.
Kelly is most useful as an upper bound: if your calculation produces a Kelly percentage of 15%, trading more than 15% per trade is provably suboptimal even in the theoretical case. If you are currently risking 2% and Kelly says 15%, you have room to increase size. If Kelly says 3% and you are risking 2%, you are close to the theoretical optimum already.
Anti-Martingale Systems
An anti-martingale approach increases position size after wins and decreases it after losses - the opposite of the martingale system, which increases after losses. The logic is to press winning streaks and pull back during losing ones, letting profitable runs compound while limiting the damage from drawdown periods.
Practical implementations include:
- Fixed ratio: A new lot size unit is added after the account gains a fixed dollar amount (delta). The delta typically increases as account size grows, making position increases harder to trigger at larger account sizes.
- Percent increase on win: After each winning trade, the risk percentage is stepped up by a small amount; after each losing trade, it steps back down. This creates a naturally responsive system that concentrates larger positions during hot streaks.
Anti-martingale approaches are well-suited to trend-following strategies where winning periods tend to cluster - you want to be larger when the system is performing. They are less suited to mean-reversion strategies where wins and losses are more randomly distributed.
Volatility-Adjusted Sizing
A fourth approach, used frequently by institutional systematic traders, adjusts position size based on the current volatility of the instrument. The typical implementation targets a fixed dollar risk that corresponds to a defined multiple of the instrument's Average True Range (ATR). If ATR is high, the position is smaller for the same dollar risk; if ATR is low, the position is larger.
This produces consistent risk across instruments and market regimes. A position on a pair with 80-pip average daily range is sized differently from a position on a pair with 40-pip average daily range, even if both are classified as "1% risk" trades. For multi-pair strategies, this is particularly valuable - it prevents the inadvertent over-concentration of risk in the most volatile instrument in the portfolio.
Choosing the Right Framework
For most retail traders, fixed fractional sizing between 0.5% and 2% per trade is the appropriate starting point. It is simple to implement, mathematically sound, and does not require sophisticated calculation. Start at the lower end of that range while you are establishing whether a strategy is genuinely profitable in live conditions.
Volatility-adjusted sizing is worth adding once you are comfortable with basic fixed fractional sizing and are trading multiple instruments. It removes the hidden risk differential between pairs with different volatility profiles without requiring you to manage a complex sizing model manually.
Full Kelly is an educational concept for most retail traders, not a practical implementation guide. Its value is in setting an upper bound on sensible risk levels, not in prescribing exact trade sizes.