Most protein antigens identified as malaria vaccine candidates are polymorphic in natural parasite populations. Current opinion is that a vaccine must be based on conserved regions of antigens, and if naturally acquired immune responses to these regions are only partially protective in humans, then the vaccine must create what is lacking in Nature. An alternative view is that a successful vaccine might need to be based on multiple allelic forms of an antigen. David Conway here shows that, far from being too pessimistic or impractical, this view offers positive ways to identify targets of protective immunity.