On May 3, 7:42=A0pm, "Phil Holman" <piholmanc@[EMAIL PROTECTED]
> wrote:
> <joseph.santanie...@[EMAIL PROTECTED]
> wrote in message
>
> news:88e7ee6e-b497-4f97-9547-dab01646ddce@[EMAIL PROTECTED]
> On May 3, 4:17 pm, "Phil Holman" <piholmanc@[EMAIL PROTECTED]
> wrote:
>
>
>
> > <joseph.santanie...@[EMAIL PROTECTED]
> wrote in message
>
>
>news:68e5c91e-2268-4a98-a9ea-19d953e05a15@[EMAIL PROTECTED]
> > On May 3, 3:14 am, Keith <nos...@[EMAIL PROTECTED]
> wrote:
>
> > > "Top rider under suspicion from pass****t
> > >
data"http://www.cyclingnews.com/news.php?id=3Dnews/2008/may08/may03new=
s&from...
>
> > > My money's on Valverde or Contador, these two would NEVER have
> > > gotten
> > > off so lightly in Puerto if they hadn't been Spanish...
>
> > "It is not unusual to have results of this kind, and there is no
> > concern at this moment," McQuaid told the Associated Press.
>
> > What does that mean? Unusal compared to what?
>
> > Data are usually normally distributed. At the 99% confidence level, 1%
> > will be statistically different (more than 2.5 standard deviations
> > from
> > the norm). 23 out of 2172 is approx 1%.
>
> > Phil H
>
> It's 854, not 2172. So do they just ignore those 23 guys, or do they
> nail them to the wall. Or do they toss a coin for each one?
>
> It depends if you want a system that blindly follows the rules with a
> known probability of making errors or if you want to take into
> consideration a known percentage of riders who will be outside of
> certain limits. There is always going to be a compromise between false
> accusation and dopers going undetected.
>
> Phil H
I know that of course, I was just pointing out how inconsistent they
are with arbitrary distinctions between when they deem something to be
im****tant and when it is not.
Joseph
|
