Why complicate an efficiency measure?
Energy delivered to the crank spider as a proportion of energy metabolised. Simple.
And why use misleading and unnecessary terms by including the word efficiency where it doesn't belong?
Crank speed variances might or might not have an influence on efficiency, but calling it an efficiency measure is totally misleading and suggests a causal link between crank speed and efficiency when none has been established (much like all these "pedal efficiency" measures falsely imply a causal link between different in-stroke torque variances and efficiency).
Just call it what it is, i.e. crank speed variance (or peak crank speed to average crank speed ratio, or whatever speed variance measure you like), and then test to see if it has any bearing on actual efficiency, or on performance.
Ironically, non-round rings are specifically designed to vary the crank speed through a pedal stroke, yet are touted as improving pedalling.
Efficiency is nice and all, but at end of the day, if you can't deliver more power for the durations of relevance for your event and when you need it (which is what I'd call effectiveness but it's a redundant term since it's just another way of saying more power), then what's the point?