Abstract This paper provides an in-depth assessment of commodity futures on applied risk measurement. We provide a thorough empirical study on deconstructed commodity futures returns and present a novel wavelet-based portfolio strategy. First, we examine the dependence structure between commodity futures and show that it is described by different dependence regimes in the short-run, medium-run and long-run. Then, the out-of-sample portfolio study unveils that daily portfolio management is mostly driven by medium-run and long-run information. Furthermore, we also find that information inherent in long-run trends outperform the information included in short-run trends and this underlines the usefulness of the wavelet approach for portfolio management.