Glenn
Sorry I didn't get to this earlier. Here are my thoughts:
1) The concentration of wear particles of a particular size in the oil at any time is a function of the actual wear rate, the volume of oil, and the rate of particle removal by filtering. Although the filter is a fairly complex creature, changing it's filtering function as it loads up, I think (don't know for sure, but suspect) that the largest filter rate changes occur initially after it is freshly replaced. I suspect that eventually it reaches a homeostasis, where the filtering rate is fairly constant for a relatively long period of time.
The cartridge type filter used in my engines seems to be very consistent over time and mileage, since my UOA wear numbers across samples at multiple oil mileage points have always been reasonable, repeatable with 1-2 ppm and made sense. That may not be the case for other filter types.
2) You are correct, some of the wear may be due to corrosion of non-wear surfaces. However, if non-wear surfaces are experiencing oxidation and corrosion, you can bet that the wear surfaces are seeing worse chemical effects, since more fresh metal is exposed, and local temperatures are higher, increasing the amount of chemical activity. These should stabilize once the new chemistry has been incorporated into the engine.
But, your point is well taken. If "things" increase on the introduction of a new oil into an old engine, this is not necessarily indicative of higher wear. It just may be that the new chemistry is doing it's thing. But, if a new equilibrium has not been achieved after the 2nd or 3rd change, and metal numbers remain higher, I've gotta argue that it's the oil. And conversely, if you introduce a new oil and the numbers go down substantially on the 1st change, something important is going on. That's why in my engines, RLI Biosyn is so amazing. There is zero doubt that wear is being reduced. The effect of the oil is near-instantaneous and easily measurable, and in all cases in recent Audi engines that I've seen shows a minimum of 2:1 (Usually 3:1) reduction in Fe measurements.
3) Yes, I think cleaning is a real phenomenon. I'd advocate that you cannot condemn an oil based on particle wear, unless it's been through 2 OCIs in the engine. On the second OCI the new chemistry has settled in. It may very well be that Redline scavenges iron oxides. I'm not sure you can say that it will smooth the surface in doing so, since bearings contain some pretty darn hard oxides. But I'm no chemist, either.
As far as the correlation to actual wear for tracking purposes, that is pretty much unassailable. There's too much published evidence. The problem with BITOG is that the analysis data is uncontrolled. There are too many labs being used, and too many different ways of contaminating the samples. I've been extremely careful with my sampling methods, and have used the same lab throughout my oil analysis projects, and I've gotten consistent, repeatable results. (I've actually taken multiple samples within short time windows, just to see whether Terry's lab is repeatable.) I'm also pretty well convinced that most labs are using the wrong equipment. ICP spectroscopy is extremely accurate, but unfortunately it has a fairly low particle size cutoff of 0.5 to 2 microns.