Position sizing determines how much capital each trade receives, making it the most consequential decision in trend following. Entry signals and market selection matter, but sizing governs drawdown depth, equity curve smoothness, and long-term survival. A robust trend system with poor sizing can fail; a mediocre system with disciplined sizing often endures long enough to capture the fat-tailed trends that drive returns.

Why Position Sizing Defines Trend Following Outcomes

Trend followers accept low win rates, often 35–45%, and rely on a handful of outsized winners. That asymmetry creates a brutal arithmetic reality: a small number of trades must pay for many small losses. If any single position is oversized, one adverse gap can erase months of gains. If positions are uniformly undersized, transaction costs and slippage consume the edge before compounding can work.

Sizing also governs the psychological durability of a system. A 20% drawdown is survivable; a 60% drawdown triggers redemptions, margin calls, and abandonment at precisely the wrong moment. Because trends cluster across asset classes, correlated positions can compound losses simultaneously. Position sizing is where correlation risk, volatility risk, and gap risk are actually managed.

Fixed Fractional Sizing: The Foundation

Fixed fractional sizing risks a constant percentage of current equity on each trade. If an account holds $500,000 and the rule is 1% risk per trade, the trade risks $5,000 regardless of the asset. Position size then derives from the distance between entry and stop:

Position Size = (Account Equity × Risk %) ÷ (Entry Price − Stop Price)

For a $50 stock with a $47 stop, the per-share risk is $3, so the position is 1,666 shares. This approach automatically shrinks exposure after losses and expands it after gains, creating an anti-martingale profile that suits trend following’s long losing streaks. It also normalizes risk across markets with wildly different volatility, from Treasury futures to cryptocurrency.

The critical discipline is that every position must have a predefined stop. Without one, “risk per trade” is undefined and sizing becomes guesswork.

Volatility-Based Sizing: ATR and Its Variants

Fixed fractional sizing depends on stop placement, which is itself a volatility question. Volatility-based sizing addresses this directly using Average True Range (ATR). A common formulation risks a fixed fraction of equity per 1 ATR of adverse movement:

Units = (Account Equity × Risk %) ÷ (N × ATR)

Where N is a multiple such as 2 or 3. A high-volatility market receives fewer units; a quiet market receives more. This equalizes the expected dollar volatility of each position, preventing a single volatile instrument from dominating portfolio risk.

Chandelier exits, trailing stops at 3 ATR, and Turtle-style 2N stops all integrate naturally with this framework. The result is a portfolio where each position contributes roughly equal risk, so no single trade determines the month’s outcome.

The Turtle Rules: A Practical Benchmark

The original Turtle experiment popularized a specific, teachable sizing system. Each market received 1 unit per $10,000 of account equity, adjusted by a volatility factor:

Unit = (1% of Account) ÷ (Market Dollar Value per ATR Point)

The system capped total exposure at 12 units, limited highly correlated markets to 6 units, and restricted single-market positions to 4 units. These caps matter as much as the base formula. They prevent the portfolio from becoming a concentrated bet on one theme—oil, interest rates, or the dollar—disguised as diversification.

Modern implementations often tighten these limits. A 2020s-era commodity portfolio might cap energy exposure at 4 units and total portfolio risk at 6% of equity, reflecting higher cross-asset correlation during crises.

Risk Parity and Inverse Volatility Weighting

Risk parity allocates capital so each position or sector contributes equal volatility rather than equal dollars. A simple implementation weights each asset inversely to its volatility:

Weight = (1 ÷ Volatility) ÷ Σ(1 ÷ Volatility)

If gold has 15% annualized volatility and 10-year notes have 5%, the bond position receives three times the capital weight of gold. The portfolio’s risk contribution from each is then similar.

For trend followers, inverse volatility weighting smooths the equity curve but can underweight the most explosive trends, which often occur in high-volatility markets. A blend—inverse volatility weighting with a floor on position size—preserves participation in outsized moves while controlling baseline risk.

Correlation-Adjusted Sizing

Trend following’s worst drawdowns occur when trends reverse simultaneously across correlated markets. Sizing must account for this. Three practical techniques help:

Cluster caps. Group markets into clusters—equity indices, energy, precious metals, rates—and limit total risk per cluster. If five equity index futures all signal long, they collectively consume one cluster budget rather than five independent allocations.

Correlation matrices. Estimate rolling correlations between open positions and scale down new positions that duplicate existing risk. This requires judgment: correlations are unstable, and crisis correlations converge toward 1.

Portfolio heat limits. Cap total open risk across all positions, often 6–10% of equity. When heat reaches the cap, new signals are skipped or sized down. This single rule prevents the most common trend-following failure: full exposure into a synchronized reversal.

Heat, Equity Curve Filters, and Drawdown Controls

Portfolio heat is the sum of open risk if all stops trigger. A trader risking 1% across ten positions carries 10% heat; a gap-driven crash could realize most of it in a day.

Equity curve filters reduce sizing during losing streaks. A common rule: when drawdown exceeds 10%, halve position sizes until equity recovers to within 5% of its high. This “equity curve trading” sacrifices some upside but dramatically reduces the probability of ruin and the depth of maximum drawdown.

Volatility targeting at the portfolio level complements this. If realized portfolio volatility exceeds a target—say 12% annualized—scale all positions down proportionally. If volatility is below target, scale up modestly, subject to heat caps.

Practical Implementation Checklist

A disciplined trend follower typically combines several layers:

• Risk 0.5–1% of equity per trade, based on stop distance
• Size positions using ATR so each carries equal dollar risk
• Cap units per market, per cluster, and per portfolio
• Limit total portfolio heat to 6–10% of equity
• Apply an equity curve filter to reduce size during drawdowns
• Recalculate sizes as equity and volatility change, not on a fixed calendar
• Account for gaps: use stops beyond obvious levels and stress-test overnight risk

Each layer addresses a distinct failure mode. Together, they convert an unpredictable stream of trend signals into a survivable, compounding portfolio.