Methods used for some time/iteration restricted improvement should receive the result that they are to improve. Currently they can only improve best.