Very true that it may just be differences between how the different P70's calibrate, as well as a lack of information on how it didn't look right.
Another thing worth pointing out that may be that we're not referring to the same 75% when we say 'calibrate at 75%' - CalMAN's advanced workflow configuration that lets you set this is referring to the luminance used, not saturation. In that case 'calibrate at 75%' (which is the workflow's default even) means 100% saturation, 75% luminance. Doing that, I could get a pretty CMS chart exactly as you posted with minimal error but the true lesser saturations (as measured on a saturation sweep, and from the various colors measured with the color checker) were horrible.
I've gotten much better results by ignoring the 100% saturation points and calibrating directly with the inner, lesser-saturated points from the saturation sweep on my set - this is part of what I mean by not using the standard calibration workflow for these sets.
If you look at my actual saturation sweep graphs, you'll see that my errors are extremely low for the inner points, drifting a bit higher on the top end (with green and yellow in particular still somewhat over-saturated), but actual total error even for them is still quite acceptable. Red of course is still an issue but bringing the red saturation sweep in-line with a hue adjustment messes up the nearby interpolated reds and oranges (i.e., skin tones
), and blue actually has a slight twist towards magenta on its top end but it's rather minor and doesn't seem to cause issues with any content.