Hi John. When you posted your analysis I posted a query about your conclusions which you never answered. As I understood your data you'd observed the relative times between froome and 'the winner' of each time trial contracting, and then concluded that this was due exclusively to Froome speeding up/increasing absolute power. I've subsequently seen you report this as established fact.
What I was questioning was the basis on which you ruled out any alternative explanations - as the very simplest example, if 'the winner's' time had on average reduced by 4.7 seconds in that period then you would expect to observe the same contraction in relative times with a 0% increase in Froome's performance. There could also be a multitude of other confounding factors in the data (for example Froome's form, motivation (not every domestique races every tt at 100%) and so on). This is not a complicated point in all honesty.
In the spirit of your user name could you respond to this peer review and explain the basis on which you have ruled out any other possible interpretation of the data set than the one that, frankly, seems to suit your agenda? Thanks.