OK, I'm back after reviewing the excellent link Citius provided. I was mistaken that the Wattbike uses a strain gauge on each crank, but it seems it doesn't and I understand now how they fudge the power balance graph between left and right leg.
However, the thread is about being clipped in and the debate, I think, started off as clipped-in being a way of generating more power by way of applying force more evenly throughout the pedal stroke and how being clipped in can help with that. A short little trip to work last night up a country lane perhaps demonstrated what I get out of being clipped in. I'm usually - no, always, clipped in. However, last night I just hopped on the bike in my work shoes, stomped on the mountain-style SPDs and pedaled. As soon as I applied a bit of power I felt my foot slip off backwards, which indicated that I do pull back at least a little bit. But that happened only once and my brain cottoned on to what was going on and I changed my pedal style and it didn't happen again without me even thinking about it.
From an engineering point of view, applying power over larger parts of the circle is a bit like a one-cyclinder engine vs a 6-cyclinder one. Whether the latter is more efficient (negating all the added friction from the many pistons) than the single pot popping up and down, I have to think about. My thinking is that the slow, pulsed power inputs waste energy trying to accelerate the large body with plenty of inertia in slow pulses rather than faster, smaller ones spaced out around the rotation.