Position sizing methods are the rules that decide how large each position is: a fixed amount, a fixed fraction of equity at risk, a size scaled to volatility, or a size derived from the Kelly criterion. The method decides how much a losing streak costs and how fast the account grows or shrinks, often more than the choice of entry signal.
How it works: four methods
- Fixed shares or fixed dollar amount. Every position is, say, 200 shares or $12,000. Simple, but the risk varies with the stop: the same $12,000 loses $480 with a stop 4% away and $1,200 with a stop 10% away. The size also ignores whether the account has grown or shrunk.
- Fixed fractional (fixed risk percentage). Each trade risks the same share of current equity, for example 1%: shares = equity × risk % ÷ (entry − stop). The loss at the stop is constant in percentage terms, positions shrink after losses and grow after gains. This is the standard method; its basics are in stop-loss and position sizing.
- Volatility-based sizing. Shares = equity × fraction ÷ ATR, so each position moves the account by a similar amount on a typical day. The Turtle traders sized their units this way; the stock version is covered in the Academy lesson stops and position sizing.
- Kelly criterion. John L. Kelly Jr. (1956) derived the fraction that maximizes long-run growth for a repeated bet with known odds. For a trade that wins R times the risk with probability W and otherwise loses the full risk: f* = W − (1 − W) ÷ R, where f* is the share of equity risked per trade.
How to read it: why full Kelly is too aggressive
Kelly's formula assumes the win rate and payoff are known exactly, every win and loss is the same size, and trades are independent. In trading, all three are estimates from a limited sample (see trading expectancy). With W = 45% and R = 2, f* = 0.45 − 0.55 ÷ 2 = 17.5% of equity per trade:
- Ten losses in a row at 17.5% leave about 15% of the account, a drawdown of about 85%. At half Kelly (8.75%) the drawdown is about 60%; at quarter Kelly (4.4%) about 36%.
- Betting more than Kelly is worse than betting less. In this example, half Kelly keeps about 76% of full Kelly's long-run growth rate, while twice Kelly cuts the growth rate to near zero.
- Estimates are noisy. If the true win rate is 40% rather than 45%, the correct fraction is 10%; risking 17.5% then yields less than half the growth the correct fraction would.
For these reasons, traders who use Kelly at all usually treat it as an upper bound and size at a fraction of it, or simply use a small fixed fraction. How much to risk per trade comes down to how deep a losing streak you can accept; the arithmetic of streaks is in max drawdown.
Worked example (illustrative numbers)
One account of $80,000, one trade: entry $50.00, stop $46.00 below support (risk $4.00 per share), ATR(20) $2.50. The system's estimated win rate is 45% with a payoff of 2.
| Method | Rule | Shares | Position | Loss at stop |
|---|---|---|---|---|
| Fixed amount | $12,000 per position | 240 | $12,000 (15%) | $960 (1.2%) |
| Fixed fractional | 1% risk: $800 ÷ $4.00 | 200 | $10,000 (12.5%) | $800 (1.0%) |
| Volatility-based | 0.5% per ATR: $400 ÷ $2.50 | 160 | $8,000 (10%) | $640 (0.8%) |
| Quarter Kelly | 4.375% risk: $3,500 ÷ $4.00 | 875 | $43,750 (54.7%) | $3,500 (4.4%) |
| Full Kelly | 17.5% risk: $14,000 ÷ $4.00 | 3,500 | $175,000 (about 2.2× the account) | $14,000 (17.5%) |
A cash account cannot even hold the full Kelly position. With a cap of 20% of equity per position ($16,000, or 320 shares), both Kelly rows are cut back to 320 shares and $1,280 of risk (1.6%): the cap, not the formula, sets the size.
Limits that apply to every method
- Maximum position size, for example 20% of equity, so a gap or a single bad report cannot dominate the account.
- Maximum total open risk: the sum of losses if every stop were hit, often called portfolio heat.
- Correlation: several stocks in one sector at 1% each behave like one larger bet.
- Liquidity: keep the position small relative to the stock's average daily dollar volume, so the stop can actually be executed near its price.
Common mistakes
- Sizing by conviction instead of by a rule.
- Plugging a short backtest's win rate into Kelly and using the full result.
- Raising the risk percentage after a winning streak, or doubling up to recover losses.
- Forgetting that a gap through the stop can cost more than the planned risk.
In Clear Tape
The free position size calculator applies the fixed-fractional method: from account size, risk percentage, entry and stop it returns the share count, dollar risk, risk per share and the position's value and share of the account. Backtests on the Systems page let you set the risk per trade from 0.1% to 5% of equity, start from $100,000 and cap any position at 20% of equity. The trade journal shows each trade's risk in $ and as a percentage of the account, and the AI period review comments on risk consistency across a week or month (Tools and Pro; see pricing).
The free Kelly criterion calculator works out full and fractional Kelly from a win rate and payoff ratio, the share of full Kelly’s growth a fraction keeps, and the caveats above.
For educational purposes only. Not investment advice.