Resistance evolution by target pests threatens the sustainability of Bt maize in Africa where insect resistance management (IRM) strategies are faced by unique challenges. The assumptions, on which current IRM strategies for stem borers are based, are not all valid for African maize stem borer species. The high dose–refuge strategy which is used to delay resistance evolution relies heavily on the presence of appropriate refuges (non-Bt plants) where pests are not under selection pressure and where sufficient numbers of Bt-susceptible individuals are produced to mate with possible survivors on the Bt maize crop. Misidentification of stem borer species and inaccurate reporting on wild host plant diversity over the past six decades created the perception that grasses will contribute to IRM strategies for these pests in Africa. Desired characteristics of refuge plants are that they should be good pest hosts, implying that larval survival is high and that it produces sufficient numbers of high-quality moths. Refuge plants should also have large cover abundance in areas where Bt maize is planted. While wild host plants may suffice in IRM strategies for polyphagous pests, this is not the case with stenophagous pests. This review discusses data of ecological studies and stem borer surveys conducted over the past decade and shows that wild host plants are unsuitable for development and survival of sufficient numbers of stem borer individuals. These grasses rather act as dead-end-trap plants and do not comply with refuge requirements of producing 500 susceptible individuals for every one resistant individual that survives on Bt maize.