Over the past 80 years there have been many advances in the development of multispecies fisheries models, from the simple predator-prey models of Vito Volterra to the highly complex ecosystem and size-based models of today. Here, we explore whether multispecies fisheries models have provided useful insight or whether they have been guilty of "muddying the waters". We conclude that multispecies models have proved useful, in particular they have shown that natural predators may take of the same order of prey fish as fishing vessels, they have provided information on the long-term prognosis for stock recovery, and they have been used to evaluate precautionary reference points for fishery management. However, the course of history has not always run smoothly, and scientists have found it difficult to describe adequately the functional relationships that exist between predators and their prey. Added to this, there are infinite ways to configure complex foodweb models, so it is often very difficult to discern the main influences on model outputs. Several promising avenues for future research are described, including individual-and size-based models, stochastic multispecies models, and coupled biogeochemical and fisheries models. One of the most promising avenues might be to provide operating models for inclusion in operational management procedures.