Recent papers have demonstrated that the yield stress and the Charpy ductile to brittle transition temperature shift at the high irradiation levels of a fusion power plant may be predicted from measurements at lower irradiation levels using neural networks. It was demonstrated that the extrapolation inherent in such predictions could be validated provided that network complexity was appropriately low. Simultaneous predictions of these metallurgical properties at the 100 dpa irradiation level and 400 degrees centigrage irradiation temperature of a possible fusion power plant have been made for a series of ferritic/martensitic steels, albeit based on mainly fission data. Together with the readily available one hundred-year activation level, benefit functions are defined which can be used to predict the most suitable alloys for a fusion power plant from within existing databases. Our model is sufficiently flexible to allow a variety of possible benefit functions to be defined. The F82H, Eurofer and LA12 alloy series all receive a favourable rating, although all results presented here must be tempered with caution until more data at relevant irradiation levels and with relevant energy spectra become available.