That's because the plot above is time vs weight, NOT one over time (1/time) vs weight. I wouldnt say that graph is linear anywhere
And yes, it would be nice to be able to investigate the behaviour under operating conditions and torques and rotations etc, but that is pretty much impossible so
@Hippo cleverly developed a benchmark test that can give an indication of a ****ed or getting-****ed VCU. its pretty much like looking at the oil in a gearbox for metal particles - the metal particles dont cause the gearbox failure but they show it is happening.
This is not infallible, for example my VCU is still working even though it gives very quick times, but the OWUT identified it was a slack VCU and even though it is working it almost certainly isnt working in the same way as a stiffer one would
At the end of the day, if the VCU is stiff at low shear rates (RPM) then it is pretty much sure to be ****ed (probably more so) at the higher RPM when you are driving, and so it is an excellent way of checking if the VCU is OK or on its way out or ****ed
A test doesnt have to give accurate values, or even nothing other than trends with no absolute values, to give very useful information