By adding a generic description of cost of resistance (COR) to the existing ‘killing-the-winner’ model, we show how this expands the model's explanatory power to include rank-abundance relationships in the host population. The model can predict a counter-intuitive relationship previously suggested in the literature, where abundant viruses are associated with rare hosts and vice versa. The model explains the observed dominance of slow-growing prokaryotes as the result of successful defence strategies, rather than as dormancy of hosts lacking essential substrates. In addition to these important conceptual aspects, the model is able to reproduce realistic values for virus : host ratios and partitioning of bacterial production between predatory loss and viral lysis. A high COR is also shown to increase the community's richness and Shannon diversity index. This model thus not only couples life strategies at the cellular level with system properties, but it also links the two system level properties of biogeochemical flows and diversity to each other. The model operates with host groups, and consequences for biodiversity when interpreting these groups in terms of species and strains are discussed.