Please make this plot for me...

[Peter Armitage]

Too busy/important to do it myself. I'd like to see...

TdF finish vs. UCI Suspicion index

UCI Rank in July 190 vs. UCI Suspicion index

[Anton Berlin]

Done !

http://www.emeraldinsight.com/content_images/fig/0050350103003.png

[Phil H]

On May 13, 11:22 am, Anton Berlin <truth_88...@yahoo.com> wrote:
> Done !
>
> http://www.emeraldinsight.com/content_images/fig/0050350103003.png

We need one of Robert Chung's charts. It'll probably look much the
same as the one you posted.
Phil H

[Fred Flintstein]

We don't have the data for the chart I'd like to see.
Which would be the UCI Suspicion Index by medical
adviser. That would tell us who is ahead of the game.

Fred Flintstein

[Simply Fred]

Fred Flintstein wrote:
> We don't have the data for the chart I'd like to see.
> Which would be the UCI Suspicion Index by medical
> adviser. That would tell us who is ahead of the game.

I wonder who does Cancellara's program (if its not Gruber).

[Fredmaster of Brainerd]

On May 13, 11:10 am, Peter Armitage <npeterarmit...@gmail.com> wrote:
> Too busy/important to do it myself.   I'd like to see...
>
> TdF finish vs. UCI Suspicion index
>
> UCI Rank in July 190 vs. UCI Suspicion index

A guy gets to be a professor, and he starts treating
everyone like a grad student.

Very well, so be it. I got your Chung Charts RIGHT HERE,
buddy. UCI suspicion index vs TDF 2010 final results:

http://mingus.as.arizona.edu/~bjw/misc/rbr/tdf2010gc_suspindex.png

I don't see much of a correlation except that the bottom
half of the GC has many more zeroes. Perhaps these
guys don't/can't want/need/afford sophisticated blood
manipulation.

Beyond the top 30 or so, no one is fighting for places,
so one might expect correlations with place number to
wash out. So maybe look at just the top 30 in the classifications:

http://mingus.as.arizona.edu/~bjw/misc/rbr/tdf2010gc_top30_suspindex.png
http://mingus.as.arizona.edu/~bjw/misc/rbr/tdf2010points_top30_suspindex.png
http://mingus.as.arizona.edu/~bjw/misc/rbr/tdf2010kom_top30_suspindex.png

Nothing dramatic leaps out. The top 5 on GC are above average
in suspicion index, but places 6-10 aren't, so you have to watch
out for cherry picking the data.

Histograms of suspicion index for top 10, 20, 50, and the
bottom half of GC:

http://mingus.as.arizona.edu/~bjw/misc/rbr/tdf2010gc_histogram_suspindex.png

I didn't do any Kolmogorov-Smirnov tests; I bet the excess
of 0-index riders in the bottom half is significant, but that the
distributions of top 10, 20, 50 are not significantly different.

Data file here:
http://mingus.as.arizona.edu/~bjw/misc/rbr/uci_susplist_tdf_2010_order1.txt

The hardest part is matching up all the name variants,
ignoring middle names and matronymics, diacritical marks etc etc.
It would be simpler if they all took one word names like
"Denilson."

Fredmaster Ben
RBR Data Mining Assistant

[Fred on a stick]

Fredmaster of Brainerd wrote:

> matronymics

Congratulations on first use in rbr.

[Brad Anders]

Excellent job, my conclusion - there's no correlation because they're
all doping. Or, there's no correlation because the UCI "suspicion
index" is bullshit. Pick your poison.

[Anton Berlin]

++ both.

[Fredmaster of Brainerd]

Also choice C - you can't make a donkey into a racehorse.
Probably all three have elements of truth. However, it doesn't
necessarily prove the suspicion index is bullshit. It's a suspicion
index, not a conviction index. One might expect it to be noisy,
and also the correlation of blood doping with results to be not 1:1.

The interesting pieces of this comparison, I think, are that the
correlation is bad (so it's not like whoever does the most
grossly obvious blood packing is automatically in the top 10)
and that the bottom half of the peloton has a lot of zeroes,
which is either an artifact of how often they get tested, or
suggests they don't do as much complex blood
manipulation.

It would be a little more revealing to compare to UCI points
since a non-GC rider can have a good Tour, winning stages
etc, and still be somewhere in the middle on GC.

Fredmaster Ben

[Frederick the Great]

In article
<6a1a2a8e-6e26-4ab6...@17g2000prr.googlegroups.com>,

Fredmaster of Brainerd <bjwe...@gmail.com> wrote:

> On May 14, 8:26 am, Brad Anders <pband...@gmail.com> wrote:
> > Excellent job, my conclusion - there's no correlation because they're
> > all doping. Or, there's no correlation because the UCI "suspicion
> > index" is bullshit. Pick your poison.
>
> Also choice C - you can't make a donkey into a racehorse.
> Probably all three have elements of truth. However, it doesn't
> necessarily prove the suspicion index is bullshit. It's a suspicion
> index, not a conviction index. One might expect it to be noisy,
> and also the correlation of blood doping with results to be not 1:1.

And therefore should never be published.

--
Old Fritz

[A. Dumas]

Fredmaster of Brainerd wrote:
> It's a suspicion
> index, not a conviction index.

Not necessarily (just) suspicion. Time since last test could also be an
important factor.

[Fred Flintstein]

On 5/15/2011 5:24 PM, Fredmaster of Brainerd wrote:
> It would be a little more revealing to compare to UCI points
> since a non-GC rider can have a good Tour, winning stages
> etc, and still be somewhere in the middle on GC.

A lot more revealing would be to have this data by medical
adviser. That's the one variable that is unaccounted for.

This is a great marketing tool for the people that are
the furthest ahead of the testing. I'll bet that isn't how
it was intended.

Fred Flintstein

[Vagina Gorilla]

On May 16, 9:01 am, Fred Flintstein <bob.schwa...@sbcremoveglobal.net>
wrote:

Dumbass. What's needed is a neural network diagram that shows Lance
at the center and the connections to doping. Popo is a good example
of 1 hop.

[Fred Flintstein]

Neural networks are a red herring.

Novitsky is a great example of this. He's spending $$$ on
one that links to Wiesel. I wish him luck, he'll need it.

Fred Flintstein