TapeScript
← Blog

Time-Series Momentum: The Trend Edge Proven Across 58 Futures Markets

July 9, 2026 · 7 min read

If you've ever heard "the trend is your friend" and rolled your eyes, here's the version with real evidence behind it. A landmark study called "Time Series Momentum" showed that an instrument's own past return predicts its future return — and it held across 58 different futures markets. This is one of the most cited trading studies ever written.

Who found it, and where

The authors are Tobias Moskowitz (Yale / AQR), Yao Hua Ooi, and Lasse Heje Pedersen (NYU Stern / AQR) — serious academics who also run real money. It was published in the Journal of Financial Economics in 2012.

What they actually did

They looked at 58 highly liquid futures — stock indexes, currencies, commodities, and bonds. For each one, the rule was simple: if the market's return over the past 12 months was positive, bet it keeps rising; if negative, bet it keeps falling. They found this worked for every single one of the 58 markets over the horizon they tested, and a basket of these bets across all asset classes produced strong returns that didn't depend on the stock market going up.

Why it should work

The paper's explanation: markets under-react to news at first (so trends persist for months), then eventually over-react (so very long-horizon returns partly reverse). Add in traders chasing performance and funds rebalancing slowly, and you get trends that last long enough to trade.

Does it still hold — honestly?

This one is already about futures, which is a big plus for retail futures traders. But two honest caveats: the original study runs to 2008-ish, and trend-following has had leaner stretches since; and the effect is measured over months, not minutes, so it's a swing/position approach, not a scalp. It's a real edge, not a free lunch.

Build it in TapeScript

Describe it plainly: "Each day, if the past 12-month return of the instrument is positive, hold long; if it's negative, hold short. Rebalance monthly." TapeScript will code it, backtest it with realistic costs and no lookahead on MNQ, MES, MGC or forex, and then run it through the gauntlet — walk-forward, Monte-Carlo, an untouched holdout — so you know whether the trend edge survives on your market, not just the 1985–2008 basket. You can also test shorter lookbacks (1, 3, 6 months) to see where the edge lives.

Citation: Moskowitz, T. J., Ooi, Y. H., & Pedersen, L. H. (2012). "Time Series Momentum." Journal of Financial Economics, 104(2), 228–250. Free PDF from NYU Stern or on SSRN.

Backtest the trend edge →

Frequently asked questions

What is time-series momentum?

A strategy where you look at an instrument's own past return — say the last 12 months — and bet it continues in the same direction. Research documented it across 58 futures markets in equities, currencies, commodities, and bonds.

How is it different from regular momentum?

Classic (cross-sectional) momentum buys the strongest assets and shorts the weakest relative to each other. Time-series momentum only compares an instrument to its own past — it doesn't need a basket, which makes it natural for trading a single futures contract.

What lookback period should I use?

The famous study used 12 months, but shorter windows (1, 3, 6 months) also showed momentum. The right answer depends on your market and horizon, which is why you test a range and keep the settings that stay robust rather than a single lucky number.

Put your idea through the gauntlet →