I believe
what Dan and Bob are saying (and both please correct me if I am mistaken, I'm continually trying to learn here) is that poorly managed airflow through the inlet/plenum/exit system is more detrimental than mis matched inlet size.
"Poorly managed" would include leaky seals between inlet and plenum, leaky plenum, leaky cylinder baffles, slow/turbulent exit flow, etc etc.
These will all have a greater detrimental effect on cooling efficiency and cooling drag than mismatched inlets because the drag created by inlet spillage is relatively small.
This of course assumes that you have a spill tolerant inlet shape, such as Dan's or Van's stock inlet.
The best gains are from getting the air that comes in to do the most amount of work (heat transfer) and get it out efficiently. If it is too much air (more than is necessary) you'll have a bit more drag than you need and lower temps than you need. By restricting that air mass flow (by restricting exit size), you will reduce the mass air flow to what is necessary, and have a net reduction in drag. That is because the air that is forced to spill at the inlets causes very little drag. Less than the gain you will see from having lesser, well managed, air flow through the system.
That's my theory, and after reading a lot of threads and references, talking with some fast and cool guys, and building a couple inlet plenums and exit fairings, I'm stickin to it.