Not at all. His graph matches mine, sans the poor resolution of his:
The matching frequency I use is actually what is implicitly used in Harman research. Here is a slide with different preference graphs for example:
View attachment 104247
I have circled the matching frequency I use (425 Hz). As you see, that is the common point between the graphs.
A reason why this is a more accurate method is that the fixture provides some of its highest accuracy in this region. Bass gets impacted by the seal. Higher frequencies get impacted by reflections and smallest variations. That region in 400 to 500 Hz however, is broadly reliable so makes for a very good anchor.
The method I use then is objective and repeatable. The method of having the tester eyeball the graphs and move them up and down differently for every headphone, not so much.
Net, net, you conclusion is flat wrong. There is nothing better there.