Position Sizing — The Art of Survival in Markets

Why Position Sizing Is Everything

Most traders obsess over entries and exits while ignoring the single most important variable in their trading: position size. The ranking of variables that determine long-run performance is not entry, then exit, then size. It is size, then exit discipline, then entry, then everything else. A trader with a 70% win rate can still go to zero if each trade risks too much. A trader with a 40% win rate and a 1:3 reward-to-risk profile can compound steadily for decades if every position is sized to a fixed fraction of equity. The arithmetic is unsentimental: returns compound multiplicatively, drawdowns recover asymmetrically, and ruin is absorbing.

Position sizing determines four things, in this order. It determines whether a string of normal losses ends your account or simply produces a small drawdown. It determines whether your emotional state during the trade is curious or panicked, which determines whether you honor your stop. It determines whether your edge has time to express itself across enough independent trials to matter. And it determines, mechanically, the long-run growth rate of your equity curve.

The 1-2% Rule and Why It Survives

The most widely adopted position sizing rule among professional traders is to risk no more than 1-2% of total account equity on any single trade. The persistence of this rule across multi-decade careers is not a coincidence; it is the answer that falls out of any reasonable risk-of-ruin calculation once you assume strings of correlated losses are possible.

The mechanic is simple. With a $50,000 account and a 1% per-trade risk budget, the maximum dollar loss on any one position is $500. If the trade is XYZ at $100 with a stop at $95, the per-share risk is $5, so the position is 100 shares. The notional value of the position is $10,000, which is 20% of the account; what matters is not the notional but the dollar at risk, which is $500.

Why 1% Beats 5% Even When Both Are Profitable

The right way to compare per-trade risk levels is to compute risk of ruin under realistic loss-streak assumptions. The table below assumes a system with 55% win rate and 1:1.5 reward-to-risk; it estimates the probability of a 50% drawdown and the geometric growth rate per trade.

Notice the asymmetry: doubling per-trade risk from 1% to 2% roughly doubles the geometric growth rate but increases the probability of catastrophic drawdown by a factor of six. Doubling again from 2% to 5% destroys the growth advantage entirely because the drawdowns themselves consume all the upside. This is why professional risk frameworks cluster in the 0.5-2% band; values above that band are not aggressive, they are mathematically dominated.

The Kelly Criterion

Developed by John Kelly at Bell Labs in 1956 and adapted to investment by Edward Thorp, the Kelly Criterion provides a mathematically optimal bet size that maximizes the long-run geometric growth of capital under known win probability and payoff.

If a system wins 55% of the time at 1.5:1 payoff: Kelly = 0.55 − (0.45 / 1.5) = 0.25, or 25% of equity per trade. That number is correct under the assumption that W and R are known with certainty and that returns are independent. Both assumptions fail in real trading. Win rate is estimated from a finite sample, payoff is estimated similarly, and clusters of correlated trades violate independence. The standard correction is to take a fraction of full Kelly.

Fixed Fractional vs. Fixed Ratio

Fixed Fractional

Risks a constant percentage of current equity per trade. As the account grows, dollar size grows proportionally. After losses, dollar size decreases automatically. The 1-2% rule is fixed fractional in disguise. Its strength is that it is anti-fragile to drawdowns: you cannot blow up by mechanically applying it. Its weakness is that it scales linearly with equity, which means the absolute speed of compounding is slow at the start when the equity base is small.

Fixed Ratio (Ryan Jones Method)

Ties size increases to a delta — a fixed profit amount required before adding a unit. With a $5,000 delta, you trade 1 contract until $5,000 of profit, then 2 contracts until another $10,000 of profit, then 3 contracts after another $15,000, and so on. The triangular structure means each additional unit is funded by accumulated profits rather than equity at risk.

Fixed Ratio scales more slowly initially but accelerates aggressively once profits compound. It also scales down aggressively in drawdowns, which is the desirable property: a 1-contract trader can never feel obligated to trade 4 contracts during a losing streak.

Volatility-Based Sizing (ATR Method)

The ATR method sets stops as a multiple of average true range and sizes the position so that 1R always equals the same dollar amount across instruments. The effect is that a high-volatility biotech and a stable utility carry identical dollar risk despite very different chart distances to the stop.

The ATR method is the cleanest way to make sizing decisions across a multi-name watchlist without the bias of prefering high-priced or low-priced tickers. It also forces honesty about volatility: a position that requires a $16 stop on a $200 stock is allocating real capital to absorb the noise floor, which is correct.

Portfolio Heat and Correlation

Per-trade sizing is a necessary condition for survival, not a sufficient one. Once multiple positions are open simultaneously, the right unit of risk is portfolio heat: the sum of dollar risk across all open positions, adjusted for correlation. Ten uncorrelated positions at 1% each carry 10% heat in nominal terms but somewhat less in effective terms because losses on independent trades partially diversify. Five tech longs at 2% each carry 10% heat in nominal terms but closer to 8-9% in effective terms because the positions co-move during a sector selloff.

Correlation Penalty in Practice

The cleanest practical adjustment is to multiply nominal heat by a correlation factor. Two positions in the same sector are roughly 70% correlated during a stress event; treat them as having 1.7x the standalone risk in aggregate. Two positions in different sectors but the same factor (e.g., two high-growth, high-multiple names) are roughly 50% correlated. Two positions in different sectors and different factors (e.g., a defensive utility and a high-multiple tech name) are roughly 20% correlated.

Sizing Decision Rules

Sizing is not an aesthetic decision; it is a procedure. The procedure below is the one that survives across professional desks regardless of style.

Common Mistakes

Sizing to a target P&L. Traders who size up because they want to make a specific dollar amount are sizing to greed, not to risk. Profit is the output, size is the input; reversing the relationship reliably produces drawdowns.

Holding size constant in drawdowns. Fixed fractional sizing automatically reduces dollar risk after losses; manually overriding this to keep dollar size constant is the most common path to ruin among otherwise disciplined traders.

Treating notional as risk. A $50,000 position with a $1,000 stop carries $1,000 of risk, not $50,000. Conversely, a $5,000 leveraged options position carries the full $5,000 of risk plus assignment risk if applicable. Always size to dollar risk, not notional.

Ignoring correlation. Five 1% positions in the same sector is one 4-5% position in disguise. The book may look diversified line by line and still take a single-event hit consistent with a concentrated bet.

Confusing volatility with risk. A high-vol stock is not riskier per dollar at risk if the position is sized to ATR. The ATR-sized biotech and the ATR-sized utility are equivalent risk units, even though they look different on the chart.