DannoXYZ 

The thing is most riders aren't even close to being at the level of the Group2 state-championship racers used in that study, much less the Group1 national-level riders. Also the intensities used on that study are very high at sprint & interval-type workouts. Compared to a "masher", these guys are easily 10x smoother. You'll notice from the force-diagrams that there's very little negative torque on either sides of the crank, thus detracting very little force from the leg on the power-stroke.

Even if they're not creating much torque by pulling up, they are at least pulling up enough to unweight the leg from the pedals; it's lifting its own weight and not requiring an upward-push by the opposite leg. If you can ride one-legged, you'll know that each leg is at least carrying its own weight and not sucking power from the other side. In contrast, mashers here will require substantial help from the opposite leg on the upstroke.

Your assertion that mashers don't have to work on their pedal strokes at all because they have the same pedal-stroke as racers with 5+ years of experience makes assumptions that we don't have data on. Such as the pedal-force profiles of people on this board. The data that we can compare is perhaps max-RPM spinning rates. I'm pretty sure any of those state- or national-champ riders that can do a 40km TT in 51-minutes can probably spin at 130-150 easily. Mashers here have problems reaching 80-90rpms. This obviously shows that their pedal-strokes aren't as smooth as the national-class riders and can use some improvement.