Storing two values per lens is new for the 1D X and 5DIII, not available in any previous body.
So...you're one of the first! Time permitting, can I ask you to satisfy a curiosity? How does it vary across the range? Simple experiment, since the AFMA value is recorded in the EXIF and displayed on the review image with full info (at least, in previous models). Pick arbitrary ends, say -10 and +10 or 0 and +10, then shoot at a few points in the zoom range. Is there a gradual transition between the values, or an abrupt transition at some point? I suspect the former, but the manual doesn't make it clear.
Thanks for your feedback Neuro!
Footnote at the bottom of p.108 of the manual says, as you may have seen
"When shooting with the intermediate range (focal length) of a zoom lens, the AF's point of focus is corrected relative to the adjustments made for the wide-angle and telephoto ends. Even if only the wide-angle or telephoto end has been adjusted, a correction will be made automatically for the intermediate range."
This is the only explanation I could find. I take it to mean a linear(?) interpolation of the W-T adjustments.
If so, it might not be as simple for a zoom lens with a variable aperture like my EF-S 17-85 f/4-5.6 zoom. What role does the variable aperture play in computing the adjustment for the intermediate focal lengths?
After my 5D3 w/24-105 f/4L arrives in early April, I'll be able to use the W-T adjustment method and test the intermediate focal lengths. Maybe the linear interpolation assumption can be observed/not?