Parker said:
No that's no what I said at all. I'm saying you should only use relevant results. You said "The usual method of applying statistics is to apply it to the entire population", so by that token, when judging climbing speeds of the peloton we have to consider Cavendish, Kittel and Greipel just as much as Froome, Horner and Nibali. But that would be stupid - so only the data for the top riders (top 10 maybe) - those that are fully racing - are relevant, not the entire population.
Ok, but can I suggest that those trying to avoid the broom wagon may actually be pushing themselves harder than many of those in front of them?
In terms of the 'top riders', your point is well taken that we should focus on these.
However, in that case please again reconsider JV's comment about 'outliers'.
He wasn't talking about slow outliers. Given that the subject was about confirming the absence of doping, he was by definition talking about fast outliers.
In other words, if you remove the data from those going at speeds consistent with doping and at power levels consistent with doping then it looks like there is less doping.
Mmmkay, but how does that prove there is less doping?
That sounds like we are simply viewing through rose-colored glasses.
Dave.