Thad-
Think you're making it more complex than it really is.
First, there is no requirement that you split the manifold- and I'm not convinced there is significant enough gain to make it worthwhile. Not based on data, based on unsupported opinion. Just drop the throttle bodies on the manifold- set up a synchronized linkage, and you're good to start filling in the tuning tables. If the end cylinders don't get exactly the same volume fill, is it any worse than the single point induction, either carb or throttle body?
When not tuning for emissions, there is a fairly large range of acceptable air/fuel ratios that can be safely run- if a couple of cylinders aren't exactly right- well, are you sure that all 15 million cars sold by automakers each
year don't have any variation between them?
If a single oulet header, put the Ox sensor in the collector, if a dual outlet parallel two ox sensors or use just one and blow the second one off.
And i don't share your optimism that fuel and air are completly homogenized and don't seperate as it makes 2- 90 degree turns in the manifold before entering the port.
My favorite idea, as I've previously noted, is to use six port injectors in, you guesed it, the ports, and use three dryflow throttle bodies on the log. Minimal chance for separation as the heavier fuel doesn't need to turn (many) corners.