First off: Great job again, James.
As for the performance, I'm not terribly impressed with the tower. Not at the price of $5200/pair. It's not "omg, no", but I'd have expected a bit more linearity (especially in the HF) at this price. The constant directivity with the off-axis notch around the 4kHz makes me wonder if this wouldn't come off a bit "bright" in-room. (see side note 2)
The center channel shows just how hard it is to create a good coaxial speaker (something I think Kef and Gelenec corner the market on). This showing by Paradigm is pretty bad. Poor termination of the tweeter to midrange and - from the photos - it seems the flare rate has a sharp change as you leave the tweeter assembly/waveguide and enter the midrange/waveguide. Whether or not I'm seeing that correctly, the data clearly shows some issues in the coaxial.
Side note 1: Your 10deg measurement has a blip ~400Hz that doesn't show up in the others. Interesting.
Side note 2: I know you don't have the data to do the full spin for the towers but I think it would be useful for you to load the horizontal data into something like VCAD or whatever software you use and generate the ERDI from the horizontal. It would give us a clearer idea of what the in-room response is likely to be and whether or not we can EQ it. I think it would be worthwhile if you already have the data for that.