The freewheel is not "freewheeling" unless the Nr is > Free turbine speed. The problem is in some ac where during steady FIG it is not possible to get a split off as it would require the collective to be lowered to a point where the rotor would exceed its Nrlimits. If some collective was applied during FIG to keep the rotor from overspeeding and that corresponded to say 5%Tq; that represents an engine contribution during FIG, therefore the ROD during a real autorotation would be higher. It is the magnitude of the engine contribution that is important if you have to do a real engine off as it will not be there to maintain Nr if the the donk has stopped, unlike during your training PFLs.