In addition to the 25mm tires they should have tested the recommended tire size. The CLX 64 suffer a lot with wide tires
This part is the only thing he's right about -- and it's one huge reason the test is nearly useless. That "25m but really sometimes 29mm" tire stalled out a lot of wheels and at really shallow angles....and it probably had a really bad shape in addition to a bad size for stalling wheels.
But the roval clx is probably the widest wheel in the test (21/29.9) and it was second worst at 5deg. They designed it to be this wide especially to accommodate wider tires and minimize the aero hit.
Also I call BS on 35W difference between the roval and enve at 5deg yaw. Just think about it, even at 30 mph, that's crazy.
Specialized most definitely did not design the wheel for extra wide tires. For aerodynamics, the shape of it wants 23mm or so, with a decent profile. Look at the tunnel data from blather about bikes... the s-works 22mm works the best, the 24mm next best.
It's widest well back from the tire, unlike the slab-sided designs of say zipp and HED. You want the wind to flow around the leeward side and stay attached... in this case you want to try to keep that section shape looking like an aircraft wing rather than the body of a wasp.... smaller tire will do that, 25mm and up creates a hollow right behind tire on leeward side and the wheel stalls out. And the effect is magnified the further the apparent wind direction shifts from zero degrees.
Same thing happens when you go too large on a Zipp but likely the U-shape is at least somewhat less susceptible to that effect, up to a certain point in tire size.
This likely explains this BS results in this test.