Jagartrott wrote:I further refer you to any statistical handbook. Not worth discussing this anymore if you don't know how statistics work. There is no trend break in the series depicted above (apart from the drop in time gain rate around 35 km), hence 'fading' cannot be made concrete.

I genuinely don’t know when people are trolling or not anymore.

You’re arguing Froome didn’t fade, except for the last part where he relatively speaking, faded? But you can’t measure this from the maximum time gap? This is an n=1, the data set is relative time difference between two people, statistics doesn’t even play in to it, as there are none! By that logic you can’t really say Pinot is fading today, because if you start at the beginning of the tour, he’s still performing well above the mean of the peloton, and it would be disingenuous to measure from his peak relative to the rest? Insanity.

I think you have heard about p-hacking in science, where researchers ends the study once there is a statistical significant result instead of ending it at the predetermined time at which a probability value would actually be the result of a real effect.