And a wrong model. There're a lot of calculators out there, e.g. [1]. Some more, some less sophisticated. But they all provide rules by thumb at best.
It starts at the mechanical losses, rolling resitance and aerodynamics (it's often neglected on calculations regarding climbs). Does his model take wind into account?
Perhaps measuring direct inputs would be better.
I notice that the kreuzotter model was tested with SRM cranks. I wonder if a universally applied power meter for all riders would work? Somewhat akin to the Dimension Data GPS trackers attached to the rider's seats, this year. If everyone has one, then it is an even playing field.
The problem is getting an accurate strain gauge that would work for different bikes, pedals, shoes. Plus, be consistent enough to not yield false data, given weather, crashes, mechanicals, and whatnot.
I find it ironic that both Sallet and Armstrong benefit from doubt on Froome's performance - as it validates their own agendas.
Though, perhaps Sallet got it right with 1) The two leaders in the tour grew up training at high elevations. Perhaps, the next wave of cheating will involved epigenetics. I expect the 2035 TdF to be highly suspect.
[Edit spell correction]