Our work shows that repeated, de novo mutation in ordinary genes is a viable strategy to traverse changing environments, and suggests pathogen evolution may be oversimplified without frequent sequencing to resolve reversions.