I think one can only try to calculate the RANGE for utilized energy for a packet (but just for non-related packet)... if you go in terms of data, the count of assumptions of successful delivery, network availability, and many more features would hinder the correct judgement