Trend Following in Commodities: A Proven Approach

The Core Philosophy: Profiting from Inefficiency

Trend following in commodities operates on a single, powerful premise: markets do not instantly adjust to fair value. Instead, prices move in sustained, directional trends driven by the slow diffusion of information, shifts in supply-demand dynamics, geopolitical shocks, and the herding behavior of market participants. This approach rejects the efficient market hypothesis, instead embracing the reality that momentum, once established, tends to persist longer than most traders anticipate.

The strategy is purely systematic. It does not forecast prices, analyze fundamentals, or predict news events. The only data point that matters is price itself. By identifying and riding established trends—both up and down—traders capture the large, asymmetric moves that occur in commodity cycles. The key is not to be right about the direction, but to be right about the persistence of the move once it is underway.

Why Commodities Are the Ideal Vehicle for Trend Following

Commodities possess structural characteristics that make them uniquely suited for trend-following strategies.

1. Long-Term Secular Cycles: Commodities operate in multi-year super-cycles driven by underinvestment in production, technological shifts, and macro-economic demand. The 2000s commodity super-cycle, fueled by Chinese industrialization, saw crude oil rise from $10 to $147. Trend following captured the majority of that move.

2. Supply Inelasticity: When demand surges, commodity producers cannot instantly increase output. Building a copper mine takes a decade; drilling a new oil well takes months. This supply lag creates extended price trends. Conversely, when demand collapses, production cannot be shut down overnight, leading to prolonged downtrends.

3. High Volatility and Non-Normal Distributions: Commodities exhibit fat-tailed statistical distributions. Extreme moves—both up and down—happen far more frequently than in equities or bonds. Trend followers thrive on these dislocations, as they provide the large, profitable swings that make the strategy work.

4. Noise vs. Signal: Commodity prices experience sharp, emotional reactions to headlines. The trend follower avoids the noise, focusing on the underlying signal: the direction of the moving average. This removes the psychological burden of trading volatility.

The Mechanics of a Robust Systematic Approach

A successful trend-following system in commodities is built on three pillars: entry logic, position sizing, and risk management.

Entry Logic: Channel Breakouts and Moving Averages

The most common entry signal is the Donchian Channel breakout. A buy signal occurs when price exceeds the highest closing price of the last 20 to 60 bars. A sell signal occurs when price breaks below the lowest low of that same period. This simple mechanism ensures the trader enters only after a trend is statistically confirmed.

A more refined approach uses multiple moving averages. A fast average (e.g., 20-day) crossing above a slow average (e.g., 100-day) triggers a long bias. The crossover reduces whipsaws during choppy, sideways markets.

Position Sizing: Risk Parity and Volatility Normalization

The most common flaw among novice trend followers is over-sizing. Commodities are not all equal; crude oil is far more volatile than gold. To maintain consistent risk, position sizes must be normalized by volatility.

The standard method is the Volatility-Adjusted Position Sizing formula:

• Determine the contract’s Average True Range (ATR) over the last 20 days.
• Define a fixed risk percentage of the portfolio per trade (typically 0.5% to 1%).
• Position Size = (Account Risk Amount) / (ATR * Contract Multiplier)

This ensures that a choppy, low-volatility market like wheat does not occupy the same capital as a volatile market like natural gas. It prevents a single losing trade from destroying the portfolio.

Exit Logic: The Trailing Stop

Profit-taking is the death of trend following. The strategy requires holding positions through drawdowns to capture the full trend. The exit is a trailing stop based on a multiple of ATR (e.g., 3x ATR from the highest high since entry) or a moving average cross.

The most common exit is the Chandelier Exit: Stop loss = Highest High (over 22 periods) – 3 * ATR. This stop moves up as the trend progresses, locking in profits while allowing the position to breathe during normal retracements.

Diversification Across Commodities: The Free Lunch

A robust trend-following portfolio must be diversified across uncorrelated commodity sectors. This reduces portfolio volatility while maintaining exposure to trending markets.

Essential Sectors:

• Energy: Crude oil, natural gas, gasoline, heating oil
• Metals: Gold, silver, copper, platinum
• Agriculture: Corn, wheat, soybeans, sugar, coffee
• Livestock: Lean hogs, live cattle
• Softs: Cotton, orange juice, cocoa

The secret is that commodities within the same sector can diverge significantly. Gold and copper, despite both being metals, have vastly different drivers. Gold is a store of value; copper is an industrial input. Diversifying across these sectors ensures that when one market is ranging, another is trending.

Managing the Psychological Reality of Drawdowns

Trend following works because it is emotionally painful. The strategy inevitably incurs long strings of small losses during periods of low volatility or false breakouts. A typical trend follower experiences a 20% to 40% drawdown multiple times per decade. This is the cost of admission.

Understanding the historical data is critical. A study by CME Group on trend following in commodities from 1960 to 2015 showed that the strategy produced an annualized return of 11.7% with a Sharpe ratio of 0.68. However, the maximum drawdown exceeded 50% in the 1970s and again during the 2013-2015 commodity bear market. The moments of greatest pain were precisely those that preceded the largest recoveries—the 2008 financial crisis drawdown was followed by a historic rally in gold and agricultural commodities.

The solution is not to avoid drawdowns, but to engineer the system to survive them. This requires a strict, rules-based approach with no discretionary override. The moment a trader intervenes to “avoid” a loss, the system stops working.

The Importance of Volatility Regime Filtering

Markets alternate between trending and choppy phases. A trend follower that trades continuously will suffer in range-bound markets. A powerful enhancement is a volatility regime filter that reduces exposure when markets are choppy.

• Trending Environment: Price moves in extended runs; ATR is expanding.
• Choppy Environment: Price oscillates; ATR is contracting; multiple false breakouts occur.

One filter uses the ADX (Average Directional Index) . When ADX is above 25, trend strength is present, and the system is fully engaged. When ADX falls below 20, the system reduces position sizes by 50% or pauses trading entirely until a new trend emerges. This avoids the death-by-a-thousand-cuts scenario.

Backtesting and Overfitting: The Hidden Pitfalls

Any trend-following system that looks perfect in backtest is likely overfitted. The goal is not to design a system that captures every wiggle in past data, but one that generalizes to unseen futures.

Critical backtesting considerations:

• Slippage and Commissions: Commodity futures have thin liquidity in some contracts. Include a realistic slippage estimate of 1 to 2 ticks per round-turn trade.
• Roll Yield: Commodities are term structures; rolling a position from one contract to another incurs a cost or benefit (backwardation vs. contango). The backtest must account for the actual roll schedule.
• Out-of-Sample Testing: Never test the system on the same period used to optimize parameters. Use the first 70% of data for development and the final 30% for walk-forward validation.

The Role of Leverage: A Double-Edged Sword

Trend following in commodities is typically executed through futures contracts, which offer significant leverage. A margin requirement of 5% to 10% means a trader can control large notional value with relatively little capital. This leverage amplifies returns and drawdowns.

Professional trend followers rarely use more than 10% to 15% of their account as initial margin. The remaining cash sits in risk-free assets like T-bills to act as a buffer against margin calls. This ensures the system survives the inevitable large drawdown without being forced to liquidate at a loss.

Real-World Application: A Portfolio Approach

A fully allocated trend-following portfolio might consist of 20 to 30 commodity futures contracts, each sized according to volatility. The allocation is rebalanced weekly to maintain target risk levels.

For example, a $1 million account might allocate:

• 10% to crude oil (high volatility, smaller contract size)
• 8% to gold (low volatility, larger contract size)
• 6% to corn (medium volatility)
• 5% to copper (medium volatility)
• 4% to sugar (high volatility, small contract size)

The remaining capital is held in cash or short-term treasuries. The system’s stop-loss is always set, and trades are executed automatically through a brokerage API or a commodity trading advisor (CTA).

Key Statistics and Historical Performance

• Annualized Return: 10% to 15% (net of fees, pre-tax)
• Maximum Drawdown: 30% to 50% over a 5-year period
• Win Rate: 30% to 40% of trades are profitable
• Profit Factor (Winning/Losing Trades): 2.0 to 3.0 (winners are significantly larger than losers)
• Average Holding Period: 30 to 90 days

These numbers underscore the asymmetry that defines trend following: losing many small trades to capture a few large winners.

The Competitive Edge: Simplicity and Adherence

The most sophisticated system is worthless without the discipline to follow it. The greatest enemy of the trend follower is the human instinct to front-run, to doubt, or to panic. The edge of the trader is not in the algorithm, but in the unwavering commitment to the rules.

Systematic execution removes emotion. Pre-programmed limit orders, hard stop-losses, and automated roll procedures eliminate the possibility of hesitation. The trader becomes an operator, not a predictor.