The problem with the way we picture ground effect lies in the way it has been taught. It is not a cushion of high pressure air below the aircraft, it is not only effective at a steady hover, and it always affects lift-producing surfaces, rotors and wings, when they come close to a surface. Ground effect is essentially a reduction in the induced drag of the airfoil, because the ground helps reduce the angle of attack needed to produce the lift.
If ground effect cannot help an autorotation, why does it show itself when we approach the runway in an airplane during a glide? If ground effect does not help at speed, why did WWII bombers limp home half a wingspan above the Channel, using one engine's power to hold themselves aloft?
The small effect it has on power and thrust (about 20%) is not apparent during the auto flair,but it is there. Physics might be confusing, but it likes to be consistent.