In addition to raw and RPD, calculate rank per instance. A line could show the mean per configuration and the critical difference from the best one.