Futures time series momentum

Checked October 2026

Time-series momentum buys a market that rose over the past year and sells one that fell. Moskowitz, Ooi and Pedersen published it in the Journal of Financial Economics in 2012. Backtoast ran it on futures. The published version failed our test, losing 0.6% a year over 101 trades.

Backtoast

You pay $99 for two full tests of your own strategy, and a pass is an offer of 30% of the profit Backtoast makes trading it with its own money.

Futures time series momentum

Each month the rule looks back twelve months. A market that's up is held long for the next month and one that's down is held short, and each market is sized to the same volatility, so a quiet bond future gets a bigger position than crude oil.

The run traded 23 futures. The first signal needed a year of history, so it covers about four years from late 2022.

Currency time series momentum is the same rule on nine currencies, and dual momentum strategy uses the same 12-month lookback to pick between stock markets.

When it works

Time-series momentum makes its money when markets trend for months, and Moskowitz, Ooi and Pedersen found it did best in extreme markets, which is why investors hold managed futures funds, most of them versions of this rule, next to stocks and bonds. It loses at turning points, when a trend reverses and the 12-month signal is slow to flip. Blending in shorter lookbacks, such as one and three months, is the usual fix, and a draft can test that before you spend a full test.

Origins

Moskowitz, Ooi and Pedersen tested 58 futures from 1985 to 2009, across stock indexes, bonds, currencies and commodities. The diversified portfolio had a Sharpe ratio above 1, so its return over cash was bigger than its yearly swings.

The run is Backtoast's, from 29 October 2026, after costs.