Additionally, one other kind of definition of energy efficiency is sometimes used, based on how much the actual process deviates from the thermodynamic limit. According to this definition, a perfect process would have a value of infinity for either its first-law or second-law efficiency; that is, the efficiency would be the tons of iron produced per amount of energy or free energy beyond the thermodynamic limit. Hence, for a perfect process, the denominators in these measures would be zero. No real process achieves the thermodynamic limit, of course, and so no real process has an infinite efficiency according to this last kind of definition.

It is also possible to use a more realistic counterpart of the (preceding) definition based on the comparison with the thermodynamic limit—namely, a comparison based on the most efficient possible process subject to a chosen time or rate constraint. This approach enables the user to compare, for example, the relative advantages and disadvantages (in energy efficiency terms) of higher-capacity but slow processes and lower-capacity but faster processes.

In practice, one very rarely encounters an explicitly stated definition of energy efficiency. Most commonly, people tend to use the very first definition, the amount of a task or product (the heating of a building for a specified time, the miles driven by a car, the tons of ore smelted, and so on) per direct unit of energy required for that task. When a different definition is being used, the user generally specifies which definition is being used. In this report, because the data have been taken from a very wide variety of sources, virtually none of which specified a definition, the panel assumed that the first and simplest definition was intended. This is not to imply that if the panel itself were to derive the efficiencies from primary data that it would use that same definition. The pragmatic course was taken here to allow the analysis to be carried out.