Position Sizing: How to Determine How Much to Risk Per Trade

Why Position Sizing Matters More Than You Think

Most traders spend the majority of their time searching for better entries, better indicators, better signals. Very few spend equivalent time on position sizing. This is a fundamental error. As Van Tharp demonstrated in his research on trading system design (documented in his book Trade Your Way to Financial Freedom), you can take the same set of trades with the same entries and exits and produce dramatically different results -- from consistent profitability to account ruin -- simply by changing the position sizing algorithm.

The reason is mathematical. Your trading system's expectancy (average win * win rate - average loss * loss rate) determines whether you have an edge. But position sizing determines how that edge compounds over time and, critically, whether you survive the inevitable drawdowns long enough to realize it. A system with a genuine edge will still blow up an account if position sizes are too large, because a string of losses (which is statistically inevitable even in a profitable system) will deplete the account before the edge can manifest.

Consider two traders with identical entry and exit signals. Trader A risks 1% of account equity per trade. Trader B risks 10% per trade. After a streak of 5 consecutive losses (which happens regularly even in systems with a 60% win rate), Trader A is down approximately 4.9%. Trader B is down approximately 41%. Trader A needs a 5.2% gain to recover. Trader B needs a 69.5% gain. Same signals, same market, but one trader is comfortably positioned to continue trading while the other is facing potential account ruin.

The Fixed Fractional Method

The fixed fractional method is the most widely used position sizing approach among professional traders and is the foundation upon which most other methods are built. The core principle is simple: risk a fixed percentage of your current account equity on every trade.

The Formula

The basic calculation has three inputs:

1. Account equity: your current account value
2. Risk percentage: the maximum percentage of equity you are willing to lose on this trade (typically 0.5% to 2%)
3. Per-share risk: the difference between your entry price and your stop loss price

The formula is:

A Worked Example

Suppose you have a $100,000 trading account and you risk 1% per trade. You want to buy a stock at $50.00 with a stop loss at $48.00.

• Account equity: $100,000
• Risk per trade: 1% = $1,000
• Entry price: $50.00
• Stop loss: $48.00
• Per-share risk: $50.00 - $48.00 = $2.00
• Position size: $1,000 / $2.00 = 500 shares
• Total position value: 500 * $50.00 = $25,000 (25% of account)

Notice that the position value ($25,000) is not the same as the risk ($1,000). You are not risking 25% of your account -- you are risking 1%. The position is $25,000 because your stop loss is relatively tight ($2 below entry). If your stop loss were $5 below entry ($45.00), the position would be smaller: $1,000 / $5.00 = 200 shares = $10,000 (10% of account). The fixed fractional method automatically adjusts position size based on the distance to the stop loss, ensuring that the dollar risk is constant regardless of the trade setup.

Why 0.5% to 2%?

The range of 0.5% to 2% risk per trade is not arbitrary. It comes from the mathematics of drawdowns. Even the best trading systems experience losing streaks. A system with a 60% win rate will, on average, experience a streak of 7 or more consecutive losses over the course of 1,000 trades (this is a straightforward probability calculation).

These are calculated using compound losses: (1 - risk%)^10. At 2% risk per trade, 10 consecutive losses produce an 18.3% drawdown. This is painful but recoverable -- you need a 22.4% gain to get back to breakeven. At 5% per trade, 10 losses create a 40.1% drawdown requiring a 66.9% gain to recover. At 10% per trade, the drawdown is 65.1%, requiring a 186.5% gain to recover. The relationship between drawdown and recovery is non-linear and punishing: the deeper the hole, the exponentially harder it is to climb out.

ATR-Based Position Sizing

Average True Range (ATR) is a volatility measure developed by J. Welles Wilder Jr. that captures the typical daily price range of a security. ATR-based position sizing uses this volatility measure to normalize risk across positions of varying volatility.

The Formula

Where N is a multiplier (commonly 1.5 to 3) that determines how many ATR units your stop loss is set away from your entry. The ATR is typically calculated over 14 periods (the standard default introduced by Wilder).

Why ATR-Based Sizing Is Powerful

The key advantage of ATR-based sizing is volatility normalization. Consider two stocks: Stock A has an ATR of $1.00 (low volatility) and Stock B has an ATR of $5.00 (high volatility). If you use a 2x ATR stop on both:

• Stock A: Stop = 2 * $1.00 = $2.00 from entry. With $1,000 risk: 500 shares.
• Stock B: Stop = 2 * $5.00 = $10.00 from entry. With $1,000 risk: 100 shares.

You hold 5x more shares of the low-volatility stock, but each position risks the same $1,000. The probability of getting stopped out is roughly equivalent for both positions because the stop is calibrated to each stock's own normal price movement. You are taking equal risk per position regardless of the underlying volatility.

This is the same principle behind the Turtle trading system popularized by Richard Dennis and William Eckhardt. The original Turtles used a 20-day ATR (which they called "N") and sized positions so that each position represented an equal amount of portfolio volatility. This approach prevented a single volatile stock from dominating the portfolio's risk profile.

The Kelly Criterion

The Kelly criterion, developed by John L. Kelly Jr. in a 1956 paper published in the Bell System Technical Journal, calculates the theoretically optimal fraction of your bankroll to risk on a bet in order to maximize the long-term geometric growth rate of your wealth. The original formula, designed for bets with binary outcomes (win or lose), is:

An Example

Suppose your trading system has a 55% win rate (p = 0.55, q = 0.45) and wins are on average 1.5x the size of losses (b = 1.5).

f* = (1.5 * 0.55 - 0.45) / 1.5 = (0.825 - 0.45) / 1.5 = 0.375 / 1.5 = 0.25

The Kelly criterion says to risk 25% of your account on each trade. This would maximize long-term geometric growth under the (unrealistic) assumption that you know the exact win rate and payoff ratio.

The Problem with Full Kelly

In practice, full Kelly sizing is dangerously aggressive for several reasons:

• Parameter uncertainty: You do not know your true win rate and payoff ratio with precision. They are estimated from historical data and subject to change as market conditions shift. Even small errors in these estimates can lead to dramatically different Kelly fractions.
• Extreme drawdowns: Full Kelly sizing produces very large drawdowns. Simulations show that even when the parameters are known exactly, full Kelly produces peak-to-trough drawdowns exceeding 50% in most multi-year simulations.
• Non-binary outcomes: Real trades do not have binary outcomes. You can lose varying amounts depending on slippage, gaps, and partial fills. The basic Kelly formula does not account for this distribution of outcomes.

Fractional Kelly

For these reasons, most practitioners who use Kelly-based sizing apply a fractional Kelly approach, typically one-quarter to one-half of the full Kelly fraction. In the example above, a quarter-Kelly approach would risk 0.25 * 25% = 6.25% per trade, and a half-Kelly would risk 12.5%. These fractions sacrifice some long-term growth rate in exchange for dramatically reduced drawdowns and greater robustness to parameter estimation errors.

Edward Thorp, who successfully applied the Kelly criterion to both blackjack (in his 1962 book Beat the Dealer) and the financial markets through his hedge fund Princeton Newport Partners, has advocated for half-Kelly or less in practice. The reduction in long-term growth rate from using half-Kelly instead of full Kelly is relatively modest (about 25% lower growth rate), but the reduction in maximum drawdown is substantial.

The Fixed Ratio Method

The fixed ratio method, developed by Ryan Jones and described in his book The Trading Game, takes a different approach to position sizing. Rather than risking a fixed percentage per trade, the fixed ratio method increases position size based on accumulated profits according to a predetermined "delta" parameter.

The core idea is that you increase your position size by one unit (for example, from 1 contract to 2 contracts) only after earning a certain dollar amount (the delta) per existing contract. If your delta is $5,000 and you trade 1 contract, you increase to 2 contracts after earning $5,000. To move from 2 contracts to 3, you need to earn another $10,000 (2 contracts * $5,000 delta). From 3 to 4 requires $15,000 more, and so on.

This creates a position sizing schedule where the dollar amount required between each increase grows linearly, producing a more conservative scaling of risk compared to the fixed fractional method (where position sizes grow exponentially with account equity). The trade-off is slower compounding in exchange for greater stability.

The Drawdown Recovery Problem

Understanding the mathematics of drawdown recovery is perhaps the single most important insight in position sizing. The relationship between drawdown depth and required recovery is non-linear:

This non-linearity is why professional risk managers obsess over drawdown control. A 20% drawdown is uncomfortable but manageable. A 50% drawdown is a catastrophe -- you need to double your remaining capital just to get back to where you started. At 75%, the situation is essentially unrecoverable for most traders. The entire purpose of disciplined position sizing is to keep drawdowns in the manageable range (typically under 20-25%) so that recovery remains realistic.

Portfolio-Level Position Sizing

Individual position sizing is necessary but not sufficient. You also need portfolio-level risk controls to manage the aggregate exposure across all open positions.

Maximum Portfolio Heat

"Portfolio heat" is the total open risk across all positions -- the sum of the dollar amount you would lose on each position if all stop losses were hit simultaneously. Professional traders typically cap portfolio heat at 6-10% of account equity. If you risk 1% per trade, this means a maximum of 6-10 simultaneous open positions. If you risk 0.5% per trade, you can hold 12-20 positions within the same heat limit.

Why does portfolio heat matter? In correlated selloffs -- such as the February 2018 VIX spike, the March 2020 COVID crash, or the interest rate shock of 2022 -- positions across different stocks tend to decline together. If you have 15 open long positions each risking 1%, a correlated drawdown that triggers all stops simultaneously would produce a 15% portfolio loss. Capping portfolio heat limits this worst-case scenario.

Sector Concentration Limits

Even with portfolio heat limits, concentration in a single sector creates hidden risk. If all your positions are technology stocks, a sector rotation out of tech will hit all of them simultaneously, and correlations will spike exactly when you do not want them to. A common guideline is to limit any single sector to no more than 25-30% of total portfolio risk.

Correlation-Adjusted Position Limits

More sophisticated approaches adjust position sizes based on the correlation between holdings. If you hold two stocks with a correlation of 0.90, they behave almost like a single position -- you have twice the nominal exposure but only slightly more than one position's worth of diversification. Some portfolio construction methods reduce individual position sizes when the portfolio's aggregate correlation is high, effectively treating highly correlated positions as a single larger position for risk purposes.

Common Position Sizing Mistakes

Sizing Based on Conviction Alone

Many traders size positions based on how strongly they believe in the trade: "I'm really confident in this one, so I'll make it a big position." This is a recipe for disaster. Conviction is not correlated with accuracy. The trades you are most confident about are often the ones where you are most biased and least objective. Disciplined position sizing should be mechanical, not emotional.

Averaging Down

Averaging down -- buying more shares as a losing position declines -- is an effective strategy in some contexts (dollar-cost averaging into index funds over years, for example), but in active trading it is almost always a mistake. When you average down, you are increasing your position size in a trade that is already going against you. This is the opposite of what sound risk management dictates.

If your original position size was calculated to risk 1% of your account, and you average down by adding another position of equal size, you have now doubled your risk to 2% on a trade that is already showing a loss. If the stock continues to decline, you hit your maximum loss sooner and with a larger position. The mathematics are unforgiving.

Ignoring Liquidity

Position sizing formulas calculate a number of shares, but that number must be feasible given the stock's liquidity. If your formula says to buy 50,000 shares of a stock that trades 100,000 shares per day, your position represents 50% of the daily volume. You will move the market against yourself entering the position and again when exiting. A reasonable guideline is to limit your position to no more than 1-5% of the average daily volume to minimize market impact.

Not Adjusting for Account Changes

Fixed fractional sizing requires recalculating position sizes based on your current account equity, not your starting equity. After a winning streak, your account is larger, and 1% of a larger account is more dollars. After a losing streak, your account is smaller, and 1% is fewer dollars. This automatic scaling -- risking more in absolute terms when you are winning and less when you are losing -- is a feature, not a bug. It is the source of the geometric compounding that makes fixed fractional sizing effective over the long term.

How Alpha Suite Handles Position Sizing

Alpha Suite's signal generation pipeline includes a built-in position sizing algorithm. For each signal, the system calculates a recommended position size using a Kelly-based framework capped at a fractional Kelly fraction of 0.25. The sizing accounts for each security's volatility (using ATR), the signal's conviction score, and the distance to the calculated stop-loss level.

The system uses a barrier model with volatility-anchored parameters to set take-profit and stop-loss levels for each signal. These levels, combined with the signal's conviction score and the configured account equity, feed into the position sizing calculation. The result is a recommended dollar allocation and share count for each signal, normalized so that higher-volatility stocks receive smaller positions and lower-volatility stocks receive larger positions -- the same ATR-based normalization principle described above.

Alpha Suite also tracks portfolio-level risk through its position management system, monitoring aggregate exposure, individual position P&L, and trailing stop adjustments. The position monitor runs multiple times during market hours to check whether any position has hit its take-profit, stop-loss, or time-stop level.