At 40 kph, rolling resistance and drive train friction combined account for ~12% of total system drag. The remaining 88% is consumed by aerodynamic drag, of which the bicycle accounts for only ~30% (in mass start trim). The rider himself is the remaining 70%.
Which makes the rider responsible for ~60% of total drag. So this alleged 1-2% improvement the equipment is responsible for necessarily must have come from only ~40% of the total friction profile.
But the chief problem with the "it
is about the bike" answer is that it assumes a 2% increase in speed equals a 2% in performance.
It doesn't
Aerodynamic drag increases with the square of the increase in velocity, but the energy required to
overcome the added drag increases
at the cube of the change. Which means any net change in rider speed occurs at the
cube root of the gross change in effort. So a 2% increase in effort, at 35 kph, will net the rider about 0.88 kph more speed. To net a 2% increase in performance would require an (2^3=) 8% increase in effort, or 8% increase in efficiency, or a combination of the two.
This chart was prepared by Ranier Pivit
Based on Pivit's data, Froome's 2013 average output was ~23% higher than Lemond's 1990 average output. 2013 was 101 km shorter than 1990 and the same number of stages, a 5 km difference average stage length.
Horner's 2013 Vuelta average output was ~24% higher than Eddy Merckx's 1973 output. 1973 was 17 stages averaging 180 km. 2013 was 21 stages averaging 160 km.
It's easy to trivialize use of rider speed to indicate doping, but less so when you understand that 2013 Horner's 3.6 kph edge over 1973 Merckx is more than simply 3.6 kph, it's also an additional 24% more energy.
Look me in the eye and tell me you find it credible that a
clean 41-year old Chris Horner could be a 24% stronger stage racer than a 28-year old Eddy Merckx.