Human–computer interface was utilized to amplify the power of a hybrid algorithm that combines fuzzy logic, regression analysis, and non-linear optimization to extract knowledge from a roll pass design database. This strategy is applied to grooves with vertical and horizontal symmetry which are employed in initial passes to roll a broad range of steel products. The results presented here from the case study are both statistically significant and valid from the viewpoint of roll pass design theory. © 2015, Springer-Verlag London.