Futures trend following

Definition

Futures trend following holds each market in the direction of its own recent trend — long what has been rising, short what has been falling — across a diversified set of futures (equity indexes, rates, currencies, commodities), with volatility-scaled positions and systematic exits. It is time-series momentum (strategy-momentum) implemented where shorting is symmetric and leverage is native.

How it works / structure

• Signals: trailing N-month excess return sign (Moskowitz-Ooi-Pedersen), moving-average crossovers, or channel breakouts (indicator-donchian-channels) — the classic variants correlate highly.
• Parameters (engine-executable): lookback(s) (commonly blended 1/3/12-month), per-market volatility target (risk-volatility-targeting with ATR or realized vol — atr_14_pct), portfolio risk cap, market universe, roll rule (ms-futures-roll — roll costs are part of every position), exit (signal flip vs trail — mgmt-stop-loss trailing variant).
• Portfolio construction: equal-risk across markets, not equal-notional; diversification across ~20+ low-correlated markets is where the strategy’s risk-adjusted profile comes from (port-diversification-math).

When it applies

As a diversified program — single-market trend following forfeits the cross-market diversification that carries the documented value; the portfolio is the strategy. Historically strongest in sustained directional regimes (including crises — “crisis alpha” behavior in 2008-type episodes per the century study) and weakest in range-bound, whipsawing regimes.

Risk profile & failure modes

• Whipsaw clusters: rangebound multi-market regimes produce long sequences of small losses; multi-year flat-to-down stretches are in the historical record and are the strategy’s cost of admission.
• Reversal gaps: trend positions are largest after long moves — sharp reversals (policy pivots, war headlines) hit maximum exposure (regime-volatility transitions).
• Roll and financing drag: carry-adverse curve shapes charge the position every cycle (ms-futures-roll, strategy-futures-carry interaction).
• Capacity and crowding: large managed-futures AUM in similar signals correlates drawdowns across the industry.

Evidence & limits

Moskowitz-Ooi-Pedersen (2012) documented time-series momentum across 58 futures markets over 1965-2009; Hurst-Ooi-Pedersen (2017) extended the evidence to a century across markets, including consistent positive performance in major equity drawdowns of their sample. Post-2009 live performance of the industry has been visibly weaker than the long-sample statistics — period-dependence is real, and the premium’s persistence is an open question the platform treats as such.

Falsifiable-thesis examples

Illustrations only, not signals:

• “A blended 1/3/12-month trend program on 20 liquid futures at 10% portfolio vol will finish the next 12 months positive in replay” — falsified by the replay P&L.
• “Crude oil’s 12-month trend sign, positive today, will remain positive at each of the next 3 month-ends” — falsified by any negative month-end reading.

Cross-references

• Signal family: strategy-momentum (time-series section), strategy-breakout
• Instrument mechanics: instrument-futures-contract, ms-futures-roll, ms-futures-margin
• Risk machinery: risk-volatility-targeting, risk-max-drawdown-budget, port-diversification-math
• Curve interaction: strategy-futures-carry