An N=2 data set doesn't prove much.
A very strict training plan doesn't mean they were doing the right things either.
Taking a natural athlete and putting them on a bike will yield good results, regardless of their training program, and some times in spite of it.
I'm no Indurain, Merckx or Coppi, but I've improved more than most simply by riding my bike ~15hrs a week, monitoring my progress with my PT, and racing my ass off.