Nice review! You know you have a fantastic PSU on your hands when just about the only things you can fault it for are a lower efficiency 5VSB and a somewhat long housing...
Perhaps the latter is due to airflow dynamics? Probably SF's testing has shown the fan to be more efficient in a bulkier casing.
You didn't mention the peculiar behavior of the PSU at low loads, with regards to AC power draw... Did you ever find out what it was all about?
I have a theory of my own... SF probably uses one of the three APFC FETs as a driver for the other two, alternating which one of them it drives based on the current/voltage phase. The On-Semi controller SF used is a CCM controller after all, and it's capable of operating in both peak and average current mode. If they used average mode (which is more likely, since this particular part works at 67kHz, too slow for ultra-precise peak-mode operation at kW power levels), that would explain phase shifting the input to accommodate for switching time (fitting rise/fall of the FETs at lower bias into the incoming AC).
Does this make any sense to you?