Ferminal said:
But it's all the same line. If the 2min AWC value is invalid then the corresponding CP value means nothing either. For a given set of data and fixed CP there is only one solution for AWC, you can't not accept that number but pretend the CP value still holds. You wouldn't use CP if you were trying to disregard the anaerobic component, there are easier ways to get a rough idea of what someone is capable of for an hour. Obviously when Froome does 6.2 in a 40km in 2011 you'd think his aerobic ability is better than what this indicates. Either that or the anaerobic component is true...
There's a big difference between some value being somewhat inaccurate and it's meaning nothing. Again, as you extrapolate to zero, the values are going to be less and less accurate, because the assumptions underlying the curve--that power equals anaerobic reserves over time--break down. E.g., there are limits to the rate at which anaerobic reserves can be tapped into. At longer intervals of time, those limits are not a factor, whereas at shorter periods they are. IOW, the curve suggests Froome has enough reserves so that if he could tap into them all at once, he could have such and such power at two minutes. But he can't necessarily mobilize them all in that period, whereas he can mobilize them, gradually, over a much longer period of time.
That does not mean that at other parts of the curve the assumptions may not be much more accurate. Even if you don't strictly buy the concept of CP, it's clear from any power curve that the decrease in power slows down as time is extended, to the point that we can estimate some value which is not much less than the value at one hour. Infinity isn't a problem here, if there isn't a genuine asymptope there is still a value at some very long period of time that can be estimated fairly accurately. And again, all riders are being treated the same. If, e.g., Horner's climbs at longer times are faster than Froome's, whereas the reverse is the case for shorter climbs, then one can definitely conclude--IF THOSE TIMES AND GRADIENTS ARE ACCURATE--that Froome has a higher anaerobic component and Horner a higher aerobic component. Even if there is no such thing as a CP.
To me, the significance of this study is that it takes into account an anaerobic component, and that different riders may have very different relative contributions from the aerobic and anaerobic components. This leads to predictions about how well they will do on short vs. long climbs, as well as, of course, speculation about what kind of doping they would be engaging in if they are doping.
And what I find most interesting, though not at all surprising, is that people get their panties all into a wad depending on where they stand on Froome and Horner. Mention the possibility that this suggests that your devil is clean and your hero is dirty, and the study becomes nonsense, automatically. Or conversely, if it supports that your hero is clean and your devil is dirty, the study is great. One can accept this method without being confined to particular conclusions about either. E.g., as I said earlier, there is a selection problem. Froome's ITT in the Vuelta suggests a higher CP than was calculated here. Also, there are no error estimates, which are particularly critical for calculating slope.