So they time him at multiple points and estimated his power from the time and the gradient at that point? Sounds very difficult to do and likely a fairly high degree of uncertainty.

And yes, if it was by time, his weight is largely irrelevant. The weight of the bike makes very little difference. About 1%.

Check it out with a power calculator. If you use 68 kg and 65 kg, with 7 kg for the bike, and a 7% gradient (I’m guessing, but doesn’t matter much), the power required for a certain constant speed is 443 watts at the heavier weight, which is 6.51 watts/kg, and 427 watts at the lower weight, which is 6.57 watts/kg. Varying the speed and gradient within the range that they would have to be doesn’t have a significant effect on the result.

So Froome's weight is essentially irrelevant to this result. Time up the climb is proportional to watts/kg, so most of the weight is already normalized. The small amount added by the bike doesn't make enough of a difference to matter.

Not to say that Froome didn't put up very high power values. YMMV, but Ross at SoS has implied that anything over 6.2-6.3 for 45 min. is suspicious. This is a shorter haul, but greater power.