I get what you are trying to say. I think that is what Ed Rosenthal was trying to say in his book. I believe what he was explaining was that let say you have a light that puts out 40,000 lumens with low PAR, and a light that puts out 5,000 lumens with high PAR, the brighter light would do a better job even though it does not put out a specific wavelength.
I agree that Ed was making the point you just made yea. But "better job" is the subjective part. I have very good understanding of the quality limitations of HIDs and I assure you PAR approach will produce the finest medicine possible (not the MOST as with many industries/processes, quality/quantity ratio is a choice not an obligation) for the least Kwatts, something that matters more to some of us than others as electricity rates vary greatly from region to region. Although I don't believe the disclaimer is 100% necessary but here it is:
assuming EVERYTHING ELSE is dialed in you should see a dramatic improvement in "quality" both medicinal compound content and terpenoid content from dialed in Red/Blue/UV/IR spectrums, which may even just involve supplementing your HID with some side-riding t5 units. There are tons of possible DIY units and daisy-chain units that you could integrate into ANY setup.
The point is that you are using hard science (PAR has lots of great research behind it) to dial in your personal results, which is hopefully what you are here for.
Not at all tryin to rip into ya, if it sounds like I'm ranting its to reiterate points for the benefit of anyone who comes across our shared work of art (thread).
Also remember this, if your area is large enough (or your reflective surface is far enough away) you could have the baddest-ass HID in the 'hood but its a point source light so its an uneven distribution with losses literally floating through the air (air-light scattering)
Also again, before you start trying to dial in your yields ANY other way, do a run in a 5gallon+ smart pot and see what you get.
Word,
MPP