Yes, I am saying those numbers are meaningless - on a DVM. A DVM works by sampling then averaging the samples, I don't know what your average DVM samples at but here's a discussion about a fairly newish DVM from Radio Shack that uses a serial port outputting at 4800bps
http://stackoverflow.com/questions/18473869/reading-data-over-serial-port-from-voltmeter
If the meter outputs at 4800 bps then it's probably got a sample rate around 4K at max - and since the meter is sampling the AC asynchronously your not going to get enough data points for the CPU in the multimeter to average them to calculate RMS voltage accurately unless your running at a frequency of perhaps no more than 1000Hz, that would be 3 data points per transition.
Of course this assumes the DVM chip is calculating RMS - my guess is in practice they run the AC into a resistor and capacitor that averages - which is worse because now you have a basic frequency filter - so your going to get voltage reading swings as frequency increases.
An interesting test would be is if you have a frequency generator to set it to generate a 3v sinewave and slowly run it from 1 Hz up to 20K Hz and see in what range the meter reads 3 volts. I'd bet there is only a narrow range of frequency it's accurate.
This kind of thing is why I still keep an analog DVM around the garage. A needle is a lot more responsive than a DVM for changing voltages.
Now I am NOT saying to throw away the voltmeter because it's still usable on COMPARITIVE measurements, that's what your missing. Meaning, you can take a perfectly working, functioning, system, connect a voltmeter, do your RPM tests, record the voltages, then take that same meter to a malfunctioning vehicle and the meter will show the fault. None of the sampling rate error matters since it's the same error introduced on both systems and is thus canceled out.
What you CAN'T do is say that on YOUR meter that at 4K rpm someone should see XYZ amount of AC voltage at the stator output because they may be using a different meter model which has different characteristics. This is why I say a scope is better - since your scope and my scope can come from different manufacturers but they will readout the same on a frequency measurement.
Now, as for the rest of it:
I agree there's a 1000 ways these things can fail and you bring up some good scenarios. We could talk about alternator failures all day long. But, limiting it to -this- forum specifically about the CB's of this year, the weak point in the system appears to be the rotor. And in my opinion, the root cause is heat, and inadequate materials selection.
I agree you can have chafing and rubbing of wires so sure, immobilization with varnish or epoxy might have some benefits. But, using cheap varnish or epoxy just means that once the rotor heats up the immobilization goes to hell because the immobilizer material disintegrates - since it can't withstand the heat.
Honda probably used the same marginal varnish on the stator coils but got away with it because the stator coils don't move. It's the old story of engineer it to be as cheap as possible so that it fails right after the warranty expires. But the rotors are clearly a design failure. Sure you can argue this is a 30 year old bike and the rotor is just old. Of course that completely ignores that the stators don't die, and it ignores that there's tons of 30 year old electric motors out there that are perfectly fine. The reality here is that Honda just used too-cheap materials. As I posted, there's magnet wire with insulation far more heat resistant, Honda should have used that wire in the rotor. I discount wire chafing as the root of the problem, I see it as more of a symptom. If the insulation isn't heat resistant then when the rotor gets hot, the insulation softens so any wire chafing or movement in the spinning rotor is going to saw right through the softened insulation like it was butter.