Doesn't the performance of non-reflecting glass
depend upon EM wave cancellation?
This I can respond to (though I doubt that I should)

The term usually used is "anti-reflective" and refers to a specific
multi-coating technique on the glass (or other medium).

the "anti" refers to destructive wave interference for reflected light
by varying the refractive index of successive coatings. Thus, MORE of
the incident travels through the medium than is reflected back

If the goal is to increase the % of light refracted THROUGH the medium,
then anti-reflective coating increases the efficiency.

Take the same glass without the AR coating, and LESS light is refracting
THROUGH the medium (more reflecting back): lower efficiency.

This definition of efficiency only makes sense when you're interested in
light refracted through the glass.

P-in / P-out does not = this definition of efficiency.

The efficiency definition implies interest in directionality.

Is this not where the confusion lies in all these posts?

John

PS,
Think of a simple flashlight with a mirror and a lens.

As a simple analogy, the filament of the light is the driven element
with a mirror behind and a lens in front...a LIGHT Yagi, no?

No diff in P-in/p-out...just all light directed out of the flashlight
into a beam.