The current decade has seen a dramatic shift in the drivers of trend-following returns. While fixed income markets were among the largest contributors to CTA performance during the 2010s, commodities have emerged as the dominant source of returns since 2020. This raises a natural question: can trend-followers improve performance by allocating more capital to sectors where trend opportunities have recently been strongest?

Our analysis suggests that the answer is largely no. Using data from 2000 through May 2026, we find that the profitability of a representative medium-term trendfollowing signal exhibits a clear horizon dependence: sectors with stronger recent trend-following performance tend to retain a relative performance advantage over subsequent months, but this effect weakens and ultimately reverses over longer horizons. Despite this predictability, reallocating capital on the basis of past trend-following performance fails to deliver a consistent improvement over a diversified trend-following portfolio.

There are two reasons why this predictability fails to translate into a robust allocation framework.

First, short-horizon Trend-on-Trend allocations are not fully independent of the underlying medium-term trend signal. Recent trend-following profitability largely reflects the same trends already embedded in the portfolio, causing the overlay to act primarily as a nonlinear amplification of existing trend exposures rather than as a distinct source of information.

Second, the strongest recent trend-following sectors tend to exhibit higher average pairwise return correlations. Reallocating toward these sectors therefore increases concentration and reduces diversification.

The location of future trends remains inherently uncertain, reinforcing a core principle of trend-following: broad diversification across markets remains more robust than attempting to time the distribution of trend opportunities across sectors.