race of truth ends tour.
Callistus Valerius
far superior to anything else out there. The Treks arn't cutting it
anymore, and how 'bout that Specialized that Levi was riding, is there
anything slower? The BMC bikes, that Phonak is riding, may be the only
thing that can compete with the Giants. Personally, I don't like Giants,
because I think they are too light, and the wheelbase is too short, but it
is obvious today they are the superior machine, in the tdf environment.
Donald Munro
Dumbass,
According to the proposed Lafferty rule everyone will be compelled to ride
a single speed track bike next year.
Donald Munro
Dumbass,
psycholist
It's not about the bike. If it were, CSC would have taken all the top
spots and Bobby Julich wouldn't have crashed 'cuz everyone knows that
Cervelos can do no wrong. At least the trithletes know that over at
Slowtwitch.com.
--
Bob C.
"Of course it hurts. The trick is not minding that it hurts."
T. E. Lawrence (of Arabia)
"Callistus Valerius" <jazz...@hotmail.com> wrote in message
news:YN_rg.5666$cd3....@newsread3.news.pas.earthlink.net...
Tim Lines
Giant has made vast improvements in their bikes this year. Last year
they were just about the same as everyone else's, this year they're
faster. The biggest improvement is that they've disposed of the
"emasculation pink" paint scheme. By my calculations, that change alone
was good for a 1:53 improvement in the stage seven time trial.
Glad someone noticed.
LastToKnow
the superior performance of almost the entire T-mobile team in the
stage 7 time trial is a worthy topic for discussion. It's a
statistical impossibility that they all "just had a good day". So
what aspect of their training, gear, or strategy accounts for their
systematically good performance?
Donald Munro
> Giant has made vast improvements in their bikes this year. Last year
> they were just about the same as everyone else's, this year they're
> faster. The biggest improvement is that they've disposed of the
> "emasculation pink" paint scheme. By my calculations, that change alone
> was good for a 1:53 improvement in the stage seven time trial.
>
> Glad someone noticed.
So there's a relationship between the pinkness of the bike and the
effectiveness of the riders scrotal patch ?
wimpyVO2
> the superior performance of almost the entire T-mobile team in the
> what aspect of their training, gear, or strategy accounts for their
> systematically good performance?
They asked Dr Fuentes to not use the phone nor to write anything down.
cycledogg
The real reason Giants look so fast this year in the Tour, Cannondales
are not there.
Tim Lines
If you've been so emasculated that you'll ride a pink bike, there is
nowhere to put the patch.
Tim Lines
> It's a
> statistical impossibility that they all "just had a good day".
I've been trying to come up with something this silly for the last 2
weeks, and you just knocked this one out of the park.
ronaldo_jeremiah
Kindly explain how it is a "statistical impossibility" that one of the
best teams in the world, riding without their original leader so that
most of their riders now have a chance to show themselves, managed to
place 1st, 4th, 6th, 8th, 14th, 16th, and 34th.
-RJ
Simon Brooke
I've said it before, and I'll say it again: either they're still getting
sprinkles on their morning cornflakes, or else they never did.
--
si...@jasmine.org.uk (Simon Brooke) http://www.jasmine.org.uk/~simon/
;; For in much wisdom is much grief; and he that increaseth
;; knowledge increaseth sorrow.." - Ecclesiastes 1:18
LastToKnow
We don't have controlled data and controlled experiments, so of course
I am speaking a bit colloquially and not the way one would do in a
scientific journal or a court of law. Nevertheless, I believe that if
one built a statistical model predicting the estimated time trial
performance (differential) of all the riders participating in TdF stage
7 based on their past performances, all of the T-mobile riders were
about a minute or so faster than what such a model would have
predicted. Like I said, I don't have such a model and I'm just
giving an opinion on Usenet, but I believe such a model could be
constructed and would come out about in the way I'm describing it. If
so, one could also put some kind of confidence bars around the
estimates, and it would be a statistical impossibility for all of those
riders to be so many standard deviations above their expected
performance in the same race.
Tim Lines
In the words of the great Vizzini: "INCONCEIVABLE!"
Ryan Cousineau
Tim Lines <SP...@SPAM.com> wrote:
You dumbass know-nothings. That's not pink, it's _magenta_.
I helped a team-mate assemble a team-colours Giant this Spring. Nice
bike.
No, I can't tell the difference between magenta and pink either,
--
Ryan Cousineau rcou...@sfu.ca http://www.wiredcola.com/
"I don't want kids who are thinking about going into mathematics
to think that they have to take drugs to succeed." -Paul Erdos
LastToKnow
The modern doping scandals never seem to involve whole teams. Given
that, the current environment of suspicion, and the fact that T-mobile
is such a well-funded team, it seems much more likely that they found
some legal means to do better in this time trial (which they will
naturally keep mum about for as long as possible).
benjo maso
"LastToKnow" <grok...@yahoo.com> wrote in message
news:1152485087.1...@35g2000cwc.googlegroups.com...
IMO the point is that the T-Mobile, probably the team with the highest
budget, has a lot of excellent riders. They were expected to help Ullrich,
but now that they are on their own, they are free to race as well as they
can. The same goes for Discovery, which also lost its leader, but there is a
big difference. Armstrong was always asking strict obedience. Anyone with a
tendence to independence - Hamilton, Landis, Boonen, etc,. - had to leave
the team. All that remained were yes-men, who this year are suddenly on
their own. No wonder they are disappointing: they are not used to be
independent. The situation in T-Mobile was quite different: Ullrich was
never able to impose his team mates the same loyalty, obedience, etc. as
Armstrong (and by the way, I exposed this theory already before the Tour,
interviewed by a Dutch bicycle magazine, explaining why Landis could win the
Tour and Hincapie never, so I was rather pleased with the results of the
TT).
Benjo
Tim Lines
> In article <zfKdnZ3rQs1H0izZ...@comcast.com>,
> Tim Lines <SP...@SPAM.com> wrote:
>
>
>>Donald Munro wrote:
>>
>>>Tim Lines wrote:
>>>
>>>
>>>
>>>>Giant has made vast improvements in their bikes this year. Last year
>>>>they were just about the same as everyone else's, this year they're
>>>>faster. The biggest improvement is that they've disposed of the
>>>>"emasculation pink" paint scheme. By my calculations, that change alone
>>>>was good for a 1:53 improvement in the stage seven time trial.
>>>>
>>>>Glad someone noticed.
>>>
>>>
>>>So there's a relationship between the pinkness of the bike and the
>>>effectiveness of the riders scrotal patch ?
>>>
>>
>>If you've been so emasculated that you'll ride a pink bike, there is
>>nowhere to put the patch.
>
>
> You dumbass know-nothings. That's not pink, it's _magenta_.
>
Which? The bike? Or my scrotum?
Patricio Carlos
Ryan Cousineau
Tim Lines <SP...@SPAM.com> wrote:
In the words of the great Vayer: "pas naturellement"
Liked Inigo Montoya the best, and especially his training regimen, which
was detailed in the book,
benjo maso
"Patricio Carlos" <pg...@hotmail.com> wrote in message
news:1152490093....@s13g2000cwa.googlegroups.com...
> So are TMO doping more? Or are DSC doping less? Or both? Please help.
As Tim Krabbé, author of the little masterpiece "The Rider", was saying the
day before yesterday on Belgian TV": "In bicycle racing there are 67
subjects more interesting than doping". Unfortunately, not everybody knows
it.
Benjo
RonSonic
>In article <zfKdnZ3rQs1H0izZ...@comcast.com>,
> Tim Lines <SP...@SPAM.com> wrote:
>
>> Donald Munro wrote:
>> > Tim Lines wrote:
>> >
>> >
>> >>Giant has made vast improvements in their bikes this year. Last year
>> >>they were just about the same as everyone else's, this year they're
>> >>faster. The biggest improvement is that they've disposed of the
>> >>"emasculation pink" paint scheme. By my calculations, that change alone
>> >>was good for a 1:53 improvement in the stage seven time trial.
>> >>
>> >>Glad someone noticed.
>> >
>> >
>> > So there's a relationship between the pinkness of the bike and the
>> > effectiveness of the riders scrotal patch ?
>> >
>>
>> If you've been so emasculated that you'll ride a pink bike, there is
>> nowhere to put the patch.
>
>You dumbass know-nothings. That's not pink, it's _magenta_.
>
>http://magentareign.com/
>
>I helped a team-mate assemble a team-colours Giant this Spring. Nice
>bike.
>
>No, I can't tell the difference between magenta and pink either,
I owned a Klein, it taught me that if you have to tell people it's magenta, it's
pink. If it really is magenta, it's still pink.
Ron
RonSonic
Thank you.
Ron
jerry in vermont
Exactly. The team has a budget of 30 mill, VERSUS (hah) I think 15
mill for the next team (QS). They are stacked. Suddenly, they find
themselves free of wacko jacko vino, and the big lump, Ulrich, and they
could get the strongest team in the world to work for them in teh TDF
and LAUNCH their career. If only they can get themselves on top of
GC....
I think they all buried themselves in the TT with good motivation. I
thin Gonchar wants to win the TDF, and I dont think he climbs that bad,
and he has plenty of experience.
Michael Press
<1152483415....@m73g2000cwd.googlegroups.com>,
"LastToKnow" <grok...@yahoo.com> wrote:
I'll offer a plausible scenario: T-Mobile wants the Team
Classification. Teams that set out to put a man on the GC
podium will not allow all team members to hammer on the
ITT. T-Mobile sent everybody out to place as highly as
possible on the ITT. Coupled with statistically outrageous
sub-par performances by other GC contenders and we have
T-Mobile 3 minutes ahead in the team classification, a
lead they will take into Paris. The yellow jersey is
lagniappe; a stick they can use to flog the other teams.
And I had Ullrich to win the GC ...
--
Michael Press
RonSonic
wrote:
I can only imagine how badly TMO must wish there were TTT this year.
Ron
b...@mambo.ucolick.org
> ronaldo_jeremiah wrote:
> > Kindly explain how it is a "statistical impossibility" that one of the
> > best teams in the world, riding without their original leader so that
> > most of their riders now have a chance to show themselves, managed to
> > place 1st, 4th, 6th, 8th, 14th, 16th, and 34th.
>
> We don't have controlled data and controlled experiments, so of course
> I am speaking a bit colloquially and not the way one would do in a
> scientific journal or a court of law. Nevertheless, I believe that if
> one built a statistical model predicting the estimated time trial
> performance (differential) of all the riders participating in TdF stage
> 7 based on their past performances, all of the T-mobile riders were
> about a minute or so faster than what such a model would have
> predicted. Like I said, I don't have such a model and I'm just
> giving an opinion on Usenet, but I believe such a model could be
> constructed and would come out about in the way I'm describing it. If
> so, one could also put some kind of confidence bars around the
> estimates, and it would be a statistical impossibility for all of those
> riders to be so many standard deviations above their expected
> performance in the same race.
Saying you believe such a model could be constructed
and throwing around "standard deviations" and "statistical
impossibility" is worthless unless you actually construct
the model, or at least provide an outline for how it would
be constructed.
The fact that Gonchar and Rogers (former/present worlds TT
champs) and Kloden (has previously TTed well in the Tour)
are up there is not at all surprising. Sinkewitz, Kessler
and Mazzoleni, maybe more of a surprise.
The results of the prologue are often a rough predictor of
ITT (and team TT) results, although you have to make allowances
for some riders, mostly sprinters and prologue specialists,
that do well in the prologue but not at longer distances.
The T-Mobile riders in question mostly did decently in the
prologue: Mazzoleni was 30th, and Kessler 101st, although
on the evidence of the first week, Kessler is in good form,
whether assisted or not. (Guerini was 138th, but if you're
going to argue about a 34th place in a TT being suspicious,
there is no point.)
BTW, Levi Leipheimer was 36th in the prologue, which
wouldn't predict that he'd completely tank the TT, but in
retrospect might have been a sign of trouble.
Mark
>
>
> IMO the point is that the T-Mobile, probably the team with the highest
> budget, has a lot of excellent riders. They were expected to help Ullrich,
> but now that they are on their own, they are free to race as well as they
> can.
Good points, but a question: How much of that budget is tied up in
Ullrich, and how much of the resources are available to support the
remaining riders?
Mark
LastToKnow
> In article
> <1152483415....@m73g2000cwd.googlegroups.com>,
> "LastToKnow" <grok...@yahoo.com> wrote:
>
> > ronaldo_jeremiah wrote:
> > > LastToKnow wrote:
> > > > Lots of sarcasm in this thread, but figuring out the explanation for
> > > > the superior performance of almost the entire T-mobile team in the
> > > > stage 7 time trial is a worthy topic for discussion. It's a
> > > > statistical impossibility that they all "just had a good day". So
> > > > what aspect of their training, gear, or strategy accounts for their
> > > > systematically good performance?
> > >
> > > Kindly explain how it is a "statistical impossibility" that one of the
> > > best teams in the world, riding without their original leader so that
> > > most of their riders now have a chance to show themselves, managed to
> > > place 1st, 4th, 6th, 8th, 14th, 16th, and 34th.
> > We don't have controlled data and controlled experiments, so of course
> > I am speaking a bit colloquially and not the way one would do in a
> > scientific journal or a court of law. Nevertheless, I believe that if
> > one built a statistical model predicting the estimated time trial
> > performance (differential) of all the riders participating in TdF stage
> > 7 based on their past performances, all of the T-mobile riders were
> > about a minute or so faster than what such a model would have
> > predicted. Like I said, I don't have such a model and I'm just
> > giving an opinion on Usenet, but I believe such a model could be
> > constructed and would come out about in the way I'm describing it. If
> > so, one could also put some kind of confidence bars around the
> > estimates, and it would be a statistical impossibility for all of those
> > riders to be so many standard deviations above their expected
> > performance in the same race.
>
> I'll offer a plausible scenario: T-Mobile wants the Team
> Classification. Teams that set out to put a man on the GC
> podium will not allow all team members to hammer on the
> ITT. T-Mobile sent everybody out to place as highly as
> possible on the ITT. Coupled with statistically outrageous
> sub-par performances by other GC contenders and we have
> T-Mobile 3 minutes ahead in the team classification, a
> lead they will take into Paris. The yellow jersey is
> lagniappe; a stick they can use to flog the other teams.
>
> And I had Ullrich to win the GC ...
Come on, Gonchar is 36 years old. There is more to this story than
than just suddenly having a longer leash.
b...@mambo.ucolick.org
> The T-Mobile riders in question mostly did decently in the
> prologue: Mazzoleni was 30th, and Kessler 101st, although
> on the evidence of the first week, Kessler is in good form,
> whether assisted or not. (Guerini was 138th, but if you're
> going to argue about a 34th place in a TT being suspicious,
> there is no point.)
That wasn't very clear. I meant that Gonchar and Rogers did
well in the prologue, and Sinkewitz and Kloden did respectably,
so it's not a huge surprise they had a good ITT. Mazzoleni was
a little further off the prologue pace and Kessler sucked,
so maybe they are a surprise. But, Kessler had a good first
week.
Once you get down past the top 20-30 or so in a long TT, there
are guys who don't have GC aspirations and are giving it 98%,
not 100%, so a rider that gives 100% can move up some places.
Ryan Cousineau
Tim Lines <SP...@SPAM.com> wrote:
Tim, I assure you I have know way of even knowing if you have a place to
put a testosterone patch, much less what colour that place may be. For
all I know, you could be the proverbial dog on the internet.
It's all about the bike,
Ryan Cousineau
Mark <remove.ma...@remove.comcast.this.net> wrote:
Unless you think you can improve a rider by paying him more, what do you
mean, Mark?
In terms of "rider support resources" for a pro, the one pricey thing
you can do is wind tunnel time. One way or another, the riders will
probably be able to get the training and the other resources they need.
Once the race starts, there's not going to be a lot of Ullrich-specific
resources that can't be redistributed to the new #1. Maybe Jan would
have a spare TT bike where Klodi might not, but that's about the limit
of it, I think.
What were you thinking of?
Ryan Cousineau
RonSonic <rons...@tampabay.rr.com> wrote:
Easy for him to say. Krabbe wrote what is almost certainly the greatest
work of fiction on the subject of cycling, and which itself constitutes
several of the cycling subjects more interesting than doping.
But it's a bit like hockey: aficionados rarely talk about the fights,
and even more rarely as a key element of a game, but it's the first
thing most outsiders notice.
Ryan Cousineau
"LastToKnow" <grok...@yahoo.com> wrote:
> Michael Press wrote:
> > In article
> > <1152483415....@m73g2000cwd.googlegroups.com>,
> > "LastToKnow" <grok...@yahoo.com> wrote:
> >
> > > ronaldo_jeremiah wrote:
> > > > LastToKnow wrote:
> > > > > Lots of sarcasm in this thread, but figuring out the explanation for
> > > > > the superior performance of almost the entire T-mobile team in the
> > > > > stage 7 time trial is a worthy topic for discussion. It's a
> > > > > statistical impossibility that they all "just had a good day". So
> > > > > what aspect of their training, gear, or strategy accounts for their
> > > > > systematically good performance?
> > > >
> > > > Kindly explain how it is a "statistical impossibility" that one of the
> > > > best teams in the world, riding without their original leader so that
> > > > most of their riders now have a chance to show themselves, managed to
> > > > place 1st, 4th, 6th, 8th, 14th, 16th, and 34th.
> C> >
> > > We don't have controlled data and controlled experiments, so of course
> > > I am speaking a bit colloquially and not the way one would do in a
> > > scientific journal or a court of law.
>
> Come on, Gonchar is 36 years old. There is more to this story than
> than just suddenly having a longer leash.
And he's a career-long TT specialist. I may have this a little wrong,
but I think he's won five Giro stages, and four of them were TTs.
T-Mobile as a group didn't have to attack, defend, or do anything at all
for this first week, Gonchar had nothing to lose by burying himself in
the work of winning this TT, and the entire GC podium from last year is
not present for one reason or another.
I wasn't picking Gonchar as a potential winner of this stage, but in
retrospect he's not a ridiculous pick. I admit to being unfamiliar with
the depth of TT specialists in this race: Zabriskie and Millar have
distinguished themselves previously.
Now, if Michael Rasmussen had won this TT, you might have a case. If
Gonchar goes on to distinguish himself in the mountains, well, then
things get interesting.
I'm skeptical about that as a possibility. I'm pretty sure that other GC
riders will eat Gonchar alive in the mountains once they start trying
hard. But I'm most interested to see what happens next, and I should not
pretend I have any special knowledge that would help me pick a favourite.
Michael Press
> Michael Press wrote:
> > I'll offer a plausible scenario: T-Mobile wants the Team
> > Classification. Teams that set out to put a man on the GC
> > podium will not allow all team members to hammer on the
> > ITT. T-Mobile sent everybody out to place as highly as
> > possible on the ITT. Coupled with statistically outrageous
> > sub-par performances by other GC contenders and we have
> > T-Mobile 3 minutes ahead in the team classification, a
> > lead they will take into Paris. The yellow jersey is
> > lagniappe; a stick they can use to flog the other teams.
> >
> > And I had Ullrich to win the GC ...
>
> Come on, Gonchar is 36 years old. There is more to this story than
> than just suddenly having a longer leash.
Gonchar was the only one to better 50 km / hour. Looks to
me like everybody else was going backwards.
The most recent long ITT in the tdf is 49 km in 2003. The
top 10 were all better than 52 km / hour.
`The second long individual time trial. 49 km, flatter
than the first long time trial and raced in cool rain on
slippy roads rather than 30-plus ° C.'
1. David Millar (Britain) 54:05
2. Tyler Hamilton (USA) +0:09
3. Lance Armstrong (USA) +0:14
4. Jan Ullrich (Germany) +0:25
5. Laszlo Bodrogi (Hun.) +0:26
6. Vjatceslav Ekimov (Rus.) +0:56
7. Victor Hugo Pena (Col.) +1:00
8. George Hincapie (USA) +1:08
9. Sylvain Chavanel (Fra.) +1:12
10. Marzio Bruseghin (Ita.) +1:26
Let see ..., yes, Ekimov was 37 in 2003.
--
Michael Press
Simon Brooke
> The fact that Gonchar and Rogers (former/present worlds TT
> champs) and Kloden (has previously TTed well in the Tour)
> are up there is not at all surprising. Sinkewitz, Kessler
> and Mazzoleni, maybe more of a surprise.
Except that T-Mo work very hard on their team time trial performance and
are probably the best team at TTT. So every member of the team is
required to be a better than competent time trialist. Mind you, this is
also true if to a slightly lesser extent of both Disco and CSC.
It's the difference in performance between T-Mo and Disco (only one T-Mo
rider was slower than the fastest Disco rider) that I find interesting
and intriguing.
--
si...@jasmine.org.uk (Simon Brooke) http://www.jasmine.org.uk/~simon/
'Victories are not solutions.'
;; John Hume, Northern Irish politician, on Radio Scotland 1/2/95
;; Nobel Peace Prize laureate 1998; few have deserved it so much
Simon Brooke
Patricio Carlos ('pg...@hotmail.com') wrote:
> So are TMO doping more? Or are DSC doping less? Or both? Please help.
My /guess/ is that T-Mo are not doping (at least, not illegally). There
will be a considerable investigative spotlight on them and on CSC.
However, seeing T-Mo's performance has not declined while others have, I
think the rather worrying conclusion has to be that T-Mo already were
the cleaner team.
--
si...@jasmine.org.uk (Simon Brooke) http://www.jasmine.org.uk/~simon/
Ring of great evil
Small one casts it into flame
Bringing rise of Men ;; gonzoron
LastToKnow
Thanks for posting numbers from this other comparably long time trial
with top rank competition. What jumps out at me is how bunched up the
leaders' times are in this other 50km time trial compared to the recent
stage 7. Gonchar didn't just win that race at 36 years old, he
crushed the rest of the field by at least a minute and the other times
fell off rapidly from there.
LastToKnow
ronaldo_jeremiah asked me to explain what I meant by "statistical
impossibility", so the above was constructed as a way of explaining
what that would mean, not as a way of convincing that I had done the
work to back that up (which I noted above that I had not). It's true
that I didn't go into details about model construction, but there are
many different statistical approaches that could be used, and
discussing them seems outside the scope of the newsgroup. A factor
analysis model, where an individual cyclist's strength is one factor is
probably the simplest starting point: c.f.
http://en.wikipedia.org/wiki/Factor_analysis If one were actually
trying to make a strong case, you would go about building models with a
variety of different statistical techniques and show that all models
pointed to the same qualitative conclusions.
> The fact that Gonchar and Rogers (former/present worlds TT
> champs) and Kloden (has previously TTed well in the Tour)
> are up there is not at all surprising. Sinkewitz, Kessler
> and Mazzoleni, maybe more of a surprise.
It's the both the quantitative and qualitative nature of the results
that is surprising - i.e. the margin of Gonchar's victory over the
entire field, the fact that 6 of the 7 T-mobile riders beat the fastest
Discovery rider, the fact that 4 of the 7 T-mobile riders beat the TT
specialist Zabriskie of CSC and the 5th best T-mobile rider was just a
few seconds behind him etc.
Scott
psycholist wrote:
> I'm sure you were fishing for this, so I'll oblige you.
>
> It's not about the bike. If it were, CSC would have taken all the top
> spots and Bobby Julich wouldn't have crashed 'cuz everyone knows that
> Cervelos can do no wrong. At least the trithletes know that over at
> Slowtwitch.com.
>
>
No, that would be Quintana Roo, not Cervelo.
Simon Brooke
Cousineau ('rcou...@sfu.ca') wrote:
> Now, if Michael Rasmussen had won this TT, you might have a case. If
> Gonchar goes on to distinguish himself in the mountains, well, then
> things get interesting.
Gonchar, as I've documented elsewhere, is about equal with
Savoldelli - 'il Falco' - in the mountains.
--
si...@jasmine.org.uk (Simon Brooke) http://www.jasmine.org.uk/~simon/
;; Quidquid latine dictum sit, altum sonatur.
Donald Munro
> Gonchar, as I've documented elsewhere, is about equal with
> Savoldelli - 'il Falco' - in the mountains.
Not when going down the other side though.
Jonathan v.d. Sluis
@gododdin.internal.jasmine.org.uk:
> in message <1152500607.0...@m73g2000cwd.googlegroups.com>,
> b...@mambo.ucolick.org ('b...@mambo.ucolick.org') wrote:
>
>> The fact that Gonchar and Rogers (former/present worlds TT
>> champs) and Kloden (has previously TTed well in the Tour)
>> are up there is not at all surprising. Sinkewitz, Kessler
>> and Mazzoleni, maybe more of a surprise.
>
> Except that T-Mo work very hard on their team time trial performance and
> are probably the best team at TTT. So every member of the team is
> required to be a better than competent time trialist. Mind you, this is
> also true if to a slightly lesser extent of both Disco and CSC.
>
> It's the difference in performance between T-Mo and Disco (only one T-Mo
> rider was slower than the fastest Disco rider) that I find interesting
> and intriguing.
>
But discovery wasn't just slower than T-Mobile. Just one of them was able
to ride in the top 20, which is a radical departure from the last couple of
years. I don't understand that. If it's connected to doping, I don't
understand what the connection is.
Curtis L. Russell
wrote:
> It's a
>statistical impossibility that they all "just had a good day".
That is simply wrong. You know, they start humming the right song on
the team bus, they get in the right mood, everything in the world
feels in harmony, pedals synchronized, they all finish high in the
standings.
Levi's wife hums some stupid song in the morning, it gets stuck in his
head, he rides like shit. I can see this.
Curtis L. Russell
Odenton, MD (USA)
Just someone on two wheels...
Curtis L. Russell
wrote:
>But it's a bit like hockey: aficionados rarely talk about the fights,
>and even more rarely as a key element of a game, but it's the first
>thing most outsiders notice.
Well, after the squid on the ice.
Jenko
Gonchar is an excellent descender as well. His good GC results in the
Giro, TTs aside, come from minimizing losses when climbing and then
reducing distances downhill. I still remember this Savoldelli-Gonchar
duo at the Giro
http://www.cyclingnews.com/results/2001/giro01/results/stage14livecomp.shtml
"1730 CEST - 147 km/16 km to go
That last kilometre really hurt. Osa starts the descent, which is very
technical. This is a Savoldelli special (but Falco is about 1 minute
behind them now).
1735 CEST - 150 km/13 km to go
Gontchar and Savoldelli are chasing down this extremely tricky descent.
The leading group have reformed (Frigo, Belli, Figueras, Simoni,
Buenahora, Contreras) and are about 10 seconds behind Uani Osa.
1738 CEST - 153 km/10 km to go
Savoldelli and Gontchar make contact with the six riders in the Maglia
Rosa group! Superb work again..."
Jenko
Donald Munro
> Gonchar is an excellent descender as well. His good GC results in the
> Giro, TTs aside, come from minimizing losses when climbing and then
> reducing distances downhill. I still remember this Savoldelli-Gonchar
> duo at the Giro
> http://www.cyclingnews.com/results/2001/giro01/results/stage14livecomp.shtml
mea culpa/dumbass. Sono testa di minchia.
I hadn't seen that, thanks.
Jenko
> Lots of sarcasm in this thread, but figuring out the explanation for
> the superior performance of almost the entire T-mobile team in the
> stage 7 time trial is a worthy topic for discussion. It's a
> statistical impossibility that they all "just had a good day". So
> what aspect of their training, gear, or strategy accounts for their
> systematically good performance?
It's something they learned from US Postal
Tour 2004
Stage 19 - July 24: Besancon - Besancon ITT, 55 km
1 Lance Armstrong (USA) US Postal p/b Berry Floor 1.06.49 (49.39 km/h)
4 Floyd Landis (USA) US Postal p/b Berry Floor 2.25
9 Jose Luis Rubiera (Spa) US Postal p/b Berry Floor 3.40
10 Jose Azevedo (Por) US Postal p/b Berry Floor 3.49
11 George Hincapie (USA) US Postal p/b Berry Floor 3.56
16 Viatcheslav Ekimov (Rus) US Postal p/b Berry Floor 4.54
27 Benjamin Noval (Spa) US Postal p/b Berry Floor 5.58
49 Manuel Beltran (Spa) US Postal p/b Berry Floor 7.26
67 Pavel Padrnos (Cze) US Postal p/b Berry Floor 8.29
Jenko
LastToKnow
> On 9 Jul 2006 10:44:13 -0700, "LastToKnow" <grok...@yahoo.com>
> wrote:
>
> > It's a
> >statistical impossibility that they all "just had a good day".
>
> That is simply wrong. You know, they start humming the right song on
> the team bus, they get in the right mood, everything in the world
> feels in harmony, pedals synchronized, they all finish high in the
> standings.
So you're familiar with some other similar situations where this sort
of thing has happened before?
> Levi's wife hums some stupid song in the morning, it gets stuck in his
> head, he rides like shit. I can see this.
It's understood that there are an overabundance of examples where an
individual turned in an inexplicably bad one day and one event
performance (Leipheimer's 2006 DL and TdF are not so dissimilar to
Mayo's 2004).
Stu Fleming
> in message <1152490093....@s13g2000cwa.googlegroups.com>,
> Patricio Carlos ('pg...@hotmail.com') wrote:
>
>
>>So are TMO doping more? Or are DSC doping less? Or both? Please help.
>
>
> My /guess/ is that T-Mo are not doping (at least, not illegally). There
> will be a considerable investigative spotlight on them and on CSC.
> However, seeing T-Mo's performance has not declined while others have, I
> think the rather worrying conclusion has to be that T-Mo already were
> the cleaner team.
>
Hence my comment about this being like 1991 (or at least, pre 1994). I
believe (and I may be wrong) that we are seeing a "non-boosting" Tour.
Maybe a "recovery products only" Tour, but without the ridiculous excess
of the EPO-era. That's also why I draw the line in the mountains at
1740 VAM.
Simon Brooke
Not getting their usual sprinkles on the cornflakes this year.
--
si...@jasmine.org.uk (Simon Brooke) http://www.jasmine.org.uk/~simon/
;; Generally Not Used
;; Except by Middle Aged Computer Scientists
Tim Lines
> In article <IZednTR2DMl3EizZ...@comcast.com>,
> Tim Lines <SP...@SPAM.com> wrote:
>
>
>>Ryan Cousineau wrote:
>>
>>>In article <zfKdnZ3rQs1H0izZ...@comcast.com>,
>>> Tim Lines <SP...@SPAM.com> wrote:
>>>
>>>
>>>
>>>>Donald Munro wrote:
>>>>
>>>>
>>>>>Tim Lines wrote:
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>>Giant has made vast improvements in their bikes this year. Last year
>>>>>>they were just about the same as everyone else's, this year they're
>>>>>>faster. The biggest improvement is that they've disposed of the
>>>>>>"emasculation pink" paint scheme. By my calculations, that change alone
>>>>>>was good for a 1:53 improvement in the stage seven time trial.
>>>>>>
>>>>>>Glad someone noticed.
>>>>>
>>>>>
>>>>>So there's a relationship between the pinkness of the bike and the
>>>>>effectiveness of the riders scrotal patch ?
>>>>>
>>>>
>>>>If you've been so emasculated that you'll ride a pink bike, there is
>>>>nowhere to put the patch.
>>>
>>>
>>>You dumbass know-nothings. That's not pink, it's _magenta_.
>>>
>>
>>Which? The bike? Or my scrotum?
>
>
> Tim, I assure you I have know way of even knowing if you have a place to
> put a testosterone patch, much less what colour that place may be. For
> all I know, you could be the proverbial dog on the internet.
>
> It's all about the bike,
>
No, I'm not a dog on the internet. Most dogs are familiar enough with
their own scrotums they don't have to ask others about them.
Hmmm ... at least I'm _pretty_ sure I'm not a dog on the internet.
ronaldo_jeremiah
Curtis L. Russell wrote:
> Levi's wife hums some stupid song in the morning, it gets stuck in his
> head, he rides like shit. I can see this.
Hums... Head...
Are you implying that Levi got a hummer from Odessa the morning of the
TT, and it sapped his 'elan vital'?
That's about the most plausible explanation I've heard so far, and
would be consistent with his modest refusal to say why he lost 6
minutes.
-RJ
ronaldo_jeremiah
> > LastToKnow wrote:
"I believe that if
> one built a statistical model predicting the estimated time trial
> performance (differential) of all the riders participating in TdF stage
> 7 based on their past performances, all of the T-mobile riders were
> about a minute or so faster than what such a model would have
> predicted."
Chung, are you reading this?
-RJ
ronaldo_jeremiah
LastToKnow wrote:
A factor
> analysis model, where an individual cyclist's strength is one factor is
> probably the simplest starting point: c.f.
> http://en.wikipedia.org/wiki/Factor_analysis
Careful, playing with tools you don't know how to use can get you hurt.
-RJ
LastToKnow
I can see why you cite this as a precedent but it is also fairly
different. Consider that this was the last competitive stage of a
tour so only some of the riders had anything at stake in terms of
either individual or team results and most were tired. It was also an
ITT with a lot of hills so only moderately decent climbers could be
competitive if they were motivated. T-mobile placed 2 riders higher
than the 2nd Discovery rider and CSC placed 3 riders higher than the
3rd Disco rider. In retrospect the only result above I would judge
as moderately surprising in relation to the competition and the
situation was Rubiera's 9th place. So I see this example as much
less anomalous or in need of explanation.
Curtis L. Russell
<ronaldo_...@yahoo.com> wrote:
>
>Hums... Head...
>
>Are you implying that Levi got a hummer from Odessa the morning of the
>TT, and it sapped his 'elan vital'?
>
>That's about the most plausible explanation I've heard so far, and
>would be consistent with his modest refusal to say why he lost 6
>minutes.
Seems like an unnecessary amount of time, considering it was the day
of the time trial. Now I'm wondering about Stephan Roche and his
problems with the TTT.
Stu Fleming
> On 10 Jul 2006 10:45:00 -0700, "ronaldo_jeremiah"
> <ronaldo_...@yahoo.com> wrote:
>
>
>>Hums... Head...
>>
>>Are you implying that Levi got a hummer from Odessa the morning of the
>>TT, and it sapped his 'elan vital'?
>>
>>That's about the most plausible explanation I've heard so far, and
>>would be consistent with his modest refusal to say why he lost 6
>>minutes.
>
>
> Seems like an unnecessary amount of time, considering it was the day
> of the time trial. Now I'm wondering about Stephan Roche and his
> problems with the TTT.
If he got one DURING the TT, it would explain the 6 minutes...
Jenko
>
> I can see why you cite this as a precedent but it is also fairly
> different. Consider that this was the last competitive stage of a
> tour so only some of the riders had anything at stake in terms of
> either individual or team results and most were tired. It was also an
> ITT with a lot of hills so only moderately decent climbers could be
> competitive if they were motivated.
The above applies to all teams, including USP.
> T-mobile placed 2 riders higher
> than the 2nd Discovery rider and CSC placed 3 riders higher than the
> 3rd Disco rider.
How does that explain USP performance?
> In retrospect the only result above I would judge
> as moderately surprising in relation to the competition and the
> situation was Rubiera's 9th place.
I see. USP had a superior team. T-Mobile has superior training.
> So I see this example as much less anomalous or in need of explanation.
So you are saying that it's not statistically impossible for a team to
place 7 riders in the Top-30 of an ITT. You just don't think T-Mobile
should be such a team.
Jenko
b...@mambo.ucolick.org
> Michael Press wrote:
> > The most recent long ITT in the tdf is 49 km in 2003. The
> > top 10 were all better than 52 km / hour.
> >
> > `The second long individual time trial. 49 km, flatter
> > than the first long time trial and raced in cool rain on
> > slippy roads rather than 30-plus ° C.'
> >
> > 1. David Millar (Britain) 54:05
> > 2. Tyler Hamilton (USA) +0:09
> > 3. Lance Armstrong (USA) +0:14
> > 4. Jan Ullrich (Germany) +0:25
> > 5. Laszlo Bodrogi (Hun.) +0:26
> > 6. Vjatceslav Ekimov (Rus.) +0:56
> > 7. Victor Hugo Pena (Col.) +1:00
> > 8. George Hincapie (USA) +1:08
> > 9. Sylvain Chavanel (Fra.) +1:12
> > 10. Marzio Bruseghin (Ita.) +1:26
> >
> > Let see ..., yes, Ekimov was 37 in 2003.
>
> Thanks for posting numbers from this other comparably long time trial
> with top rank competition. What jumps out at me is how bunched up the
> leaders' times are in this other 50km time trial compared to the recent
> stage 7. Gonchar didn't just win that race at 36 years old, he
> crushed the rest of the field by at least a minute and the other times
> fell off rapidly from there.
Look, I'm sorry, but you guys are bullshitting. You get a failing
grade on your lab report until you turn in a complete rewrite.
I see crappy data and a bad model. MIchael, the 2003 stage 19 TT to
Nantes is not the most recent long ITT in the Tour. There is almost
always a ~50 km ITT. The 2003 TT to Nantes is cherry-picking: it is
distinguished for rain and a howling tailwind that enabled riders
(juiced, not juiced, I dunno) to set a record pace. (And for Ullrich
crashing, but that is another story.) Read the cyclingnews report.
This is one reason I am flunking you, LastToKnow, for throwing around
jargon like factor analysis. To construct a decent model you need
decent measurements of the factors - like the speed of the
headwind/tailwind.
Here are the top 3 in the long ITTs from 2003-2006, results from
cyclingnews:
2003 stage 12, 47 km, very hot
1 Jan Ullrich (Ger) Team Bianchi 58.33 (48.18 km/h)
2 Lance Armstrong (USA) US Postal 1.36
3 Alexandre Vinokourov (Kaz) Telekom 2.06
2003 stage 19, 49 km, cool, rain, tailwind (record pace)
1 David Millar (GBr) Cofidis 54.05 (54.36 km/h)
2 Tyler Hamilton (USA) Team CSC 0.09
3 Lance Armstrong (USA) US Postal 0.14
2004 stage 19, 55 km, early rain, late sun
1 Lance Armstrong (USA) US Postal 1.06.49 (49.39 km/h)
2 Jan Ullrich (Ger) T-Mobile Team 1.01
3 Andreas Klöden (Ger) T-Mobile Team 1.27
2005 stage 20, 55 km, cool, Cat 3 climb
1 Lance Armstrong (USA) Discovery Channel 1.11.46 (46.4 km/h)
2 Jan Ullrich (Ger) T-Mobile Team 0.23
3 Alexandre Vinokourov (Kaz) T-Mobile Team 1.16
2006 stage 7, 52 km, mild, headwind out/tailwind back
1 Serguei Gonchar (Ukr) T-Mobile 1.01.44 (50.5 km/h)
2 Floyd Landis (USA) Phonak 1.01
3 Sebastian Lang (Ger) Gerolsteiner 1.04
So the 2006 TT was faster than all but the one to Nantes
(several of the others occurred late in the Tour and this one
is early, which probably contributes) and gaps of a minute
between first and second are not uncommon. Although they
seem a little suspicious, Gonchar probably would have beat
Landis by less than a minute if Landis had not needed a bike
change.
LastToKnow
bjw, you use the term "bullshitting" and you made two posts suggesting
you understand what a factor analysis model is and I don't. But your
statement above implying that factor analysis models should only be
used where one has a comprehensive physical description of the system
generating the data isn't even remotely related to professional
statistical practice.
Factor analysis is just one way of using and interpreting linear
models. Linear are models are the simplest, most common type of
statistical regression models which are used for everything under the
sun, especially where one doesn't know enough about the mechanisms
generating the observations or have enough a large enough data set to
use more complicated models. Moreover, the only factor I expicitly
mentioned using was the individual rider, which we obviously can use an
an explanatory variable in any linear model.
If you actually have some understanding of statistics and you think I
said something incorrect then please clarify what you are talking
about. Otherwise I think the term "bullshitting" is a very apt
description of what you wrote in response to me.
Regarding your point that 1 minute gap between 1st and 2nd place in a
50km time trial is not *by itself* especially unusual, I agree, but I
don't recall anybody saying that it was. But when the guy winning is
a 36 year old who hasn't dominated any recent races, and his entire
team does unexpectedly well in the same time trial, that is a bit more
noteworthy.
LastToKnow
> LastToKnow wrote:
> >
> > I can see why you cite this as a precedent but it is also fairly
> > different. Consider that this was the last competitive stage of a
> > tour so only some of the riders had anything at stake in terms of
> > either individual or team results and most were tired. It was also an
> > ITT with a lot of hills so only moderately decent climbers could be
> > competitive if they were motivated.
>
> The above applies to all teams, including USP.
Sure, but compared to other teams USP had
1) more good climbers
2) more guys in contention for high GC positions
3) the factor of being in contention for the team competition.
> > T-mobile placed 2 riders higher
> > than the 2nd Discovery rider and CSC placed 3 riders higher than the
> > 3rd Disco rider.
>
> How does that explain USP performance?
It explains that the gap between USP performance and performance of the
other teams was not so large in this time trial as the gap between
T-mobile and the other teams in the 2006 Stage 7.
> > In retrospect the only result above I would judge
> > as moderately surprising in relation to the competition and the
> > situation was Rubiera's 9th place.
>
> I see. USP had a superior team. T-Mobile has superior training.
You're not really trying to understand what I wrote. I said that
*all* of the T-mobile riders did unexpectedly well in the recent ITT
compared to what could reasonably be expected of them, whereas in the
2004 ITT, IMO only Rubiera did unexpectedly well. Of course you can
disagree with my opinion about those results, but you haven't made any
case that I'm applying it inconsistently.
Jenko
>
> So the 2006 TT was faster than all but the one to Nantes
> (several of the others occurred late in the Tour and this one
> is early, which probably contributes) and gaps of a minute
> between first and second are not uncommon. Although they
> seem a little suspicious, Gonchar probably would have beat
> Landis by less than a minute if Landis had not needed a bike
> change.
According to Allen Lim, Landis virtual time (without bike change and
extra drag) was 3 seconds slower than Gonchar
http://www.bicycling.com/tourdefrance/article/0,6802,s1-7-123-14884-2,00.html
Jenko
Michael Press
Curtis L. Russell <cur...@md-bicycling.org> wrote:
> On Mon, 10 Jul 2006 07:53:29 GMT, Ryan Cousineau <rcou...@sfu.ca>
> wrote:
>
> >But it's a bit like hockey: aficionados rarely talk about the fights,
> >and even more rarely as a key element of a game, but it's the first
> >thing most outsiders notice.
>
> Well, after the squid on the ice.
It's an octopus. Aficionados cook the octopus first to
make launching and clean up easier.
--
Michael Press
Jenko
>
> You're not really trying to understand what I wrote. I said that
> *all* of the T-mobile riders did unexpectedly well in the recent ITT
> compared to what could reasonably be expected of them, whereas in the
> 2004 ITT, IMO only Rubiera did unexpectedly well. Of course you can
> disagree with my opinion about those results, but you haven't made any
> case that I'm applying it inconsistently.
Basically, you are overvaluing USP'04 team (from what I remember, it was
the first Top 10 in a Tour ITT for Landis, Azevedo and Rubiera) and
undervaluing TMO'06. I personally was expecting more from Rogers (4th is
not that good a result for a 3 times world champ) and Klöden (who did
better before). Sinkewitz slightly overperformed. Mazzoleni, a
Rubiera-like rider, was the real surprise.
That's the similar situation you were asking for (and not the first one,
Once in 1994 put his 6 riders in the Top 14). If "such team has better
riders" was a valid explanation in 2004, why shouldn't it be valid now?
Jenko
b...@mambo.ucolick.org
> bjw, you use the term "bullshitting" and you made two posts suggesting
> you understand what a factor analysis model is and I don't. But your
> statement above implying that factor analysis models should only be
> used where one has a comprehensive physical description of the system
> generating the data isn't even remotely related to professional
> statistical practice.
Sorry, I was thinking of methods like principal component analysis
and used "factor" in a colloquial sense rather than in the technical
sense used in factor analysis, where it's an unobservable.
> Factor analysis is just one way of using and interpreting linear
> models. Linear are models are the simplest, most common type of
> statistical regression models which are used for everything under the
> sun, especially where one doesn't know enough about the mechanisms
> generating the observations or have enough a large enough data set to
> use more complicated models. Moreover, the only factor I expicitly
> mentioned using was the individual rider, which we obviously can use an
> an explanatory variable in any linear model.
>
> If you actually have some understanding of statistics and you think I
> said something incorrect then please clarify what you are talking
> about. Otherwise I think the term "bullshitting" is a very apt
> description of what you wrote in response to me.
In general, in linear models you need to have more constraints
than unknowns. For example, if you have known inputs and a
single output, you could construct something like
log speed = a x log power + b x log frontal area +
c x log headwind. You have say 50 riders and each of them
has a measured speed, power, area, headwind, and you do some
kind of multiple regression to find out the coefficients.
If you don't know the inputs but have multiple outputs, you can
do something like principal component analysis, or the aptitude-test
factor analysis outlined in the wikipedia page on factor analysis
that you helpfully pointed out. Here you have several measurements
for each subject and the goal (at least in PCA) is to reduce the
dimensionality of the data set. But you need more measurements
per subject than factors (hidden variables) otherwise the linear
model is underconstrained.
The problem with applying this to TTs is that there are only two
obvious known measurements, speed and distance. So you can
pile up all the TTs Kloden or Kessler ever did in his career, plot
speed versus distance, derive a mean relationship, and tell whether
stage 7 was off the relationship. Or use multiple riders to derive
the slope of the speed-distance relationship. But this isn't so
complicated that it needs to be called factor analysis.
If you want to get really fancy, you can start to introduce a per-stage
term that attempts to correlate across multiple riders to tell whether
a given TT was faster or slower than the norm - for example, suppose
75% of the riders in that crazy tailwind TT in 2003 rode faster than
usual, so the model fits a term to account for that. I'm not sure this
is factor analysis either.
Since I wasn't the one who guessed that T-Mobile were all a standard
deviation off an unspecified model, perhaps you should come up with
a better model.
BTW,
2006 Giro stage 11, 50 km ITT
1 Jan Ullrich (Ger) T-Mobile Team 58.48 (51.02 km/h)
2 Ivan Basso (Ita) Team CSC 0.28
3 Marco Pinotti (Ita) Saunier Duval-Prodir 1.01
4 Serguei Gonchar (Ukr) T-Mobile Team 1.09
2005 Giro stage 8, 45 km
1 David Zabriskie (USA) Team CSC 58.31 (46.14 km/h)
2 Ivan Basso (Ita) Team CSC 0.17
3 Paolo Savoldelli (Ita) Discovery Channel 0.44
4 Marzio Bruseghin (Ita) Fassa Bortolo 0.48
5 Serguei Gonchar (Ukr) Domina Vacanze 0.51
LastToKnow
> LastToKnow wrote:
> >
> > You're not really trying to understand what I wrote. I said that
> > *all* of the T-mobile riders did unexpectedly well in the recent ITT
> > compared to what could reasonably be expected of them, whereas in the
> > 2004 ITT, IMO only Rubiera did unexpectedly well. Of course you can
> > disagree with my opinion about those results, but you haven't made any
> > case that I'm applying it inconsistently.
>
> Basically, you are overvaluing USP'04 team (from what I remember, it was
> the first Top 10 in a Tour ITT for Landis, Azevedo and Rubiera) and
> undervaluing TMO'06. I personally was expecting more from Rogers (4th is
> not that good a result for a 3 times world champ) and Klöden (who did
> better before). Sinkewitz slightly overperformed. Mazzoleni, a
> Rubiera-like rider, was the real surprise.
As noted before, capability for the 2004 ITT was similar to capability
in a climbing stage, and both Azevedo and Landis were good climbers,
Azevedo finishing 5th overall in the GC that year based on his climbing
ability. Landis may have seemed surprising at the time because he
was relatively new to road racing, but that performance doesn't seem
surprising in retrospect. If you think we will look back in two years
on this year's T-mobile performance and not view it as either
surprising or involving factors other than the riders' comparative time
trialing strength, then I won't say you are wrong but I'm wondering why
you think that now.
> That's the similar situation you were asking for (and not the first one,
> Once in 1994 put his 6 riders in the Top 14). If "such team has better
> riders" was a valid explanation in 2004, why shouldn't it be valid now?
I don't know much about that. Were those results unexpected?
LastToKnow
> LastToKnow wrote:
>
> > bjw, you use the term "bullshitting" and you made two posts suggesting
> > you understand what a factor analysis model is and I don't. But your
> > statement above implying that factor analysis models should only be
> > used where one has a comprehensive physical description of the system
> > generating the data isn't even remotely related to professional
> > statistical practice.
>
> Sorry, I was thinking of methods like principal component analysis
> and used "factor" in a colloquial sense rather than in the technical
> sense used in factor analysis, where it's an unobservable.
There is nothing wrong with using principal components as part of
regression modeling in such a situation either...
> > Factor analysis is just one way of using and interpreting linear
> > models. Linear are models are the simplest, most common type of
> > statistical regression models which are used for everything under the
> > sun, especially where one doesn't know enough about the mechanisms
> > generating the observations or have enough a large enough data set to
> > use more complicated models. Moreover, the only factor I expicitly
> > mentioned using was the individual rider, which we obviously can use an
> > an explanatory variable in any linear model.
> >
> > If you actually have some understanding of statistics and you think I
> > said something incorrect then please clarify what you are talking
> > about. Otherwise I think the term "bullshitting" is a very apt
> > description of what you wrote in response to me.
>
> In general, in linear models you need to have more constraints
> than unknowns.
It goes without saying that in any statistical estimation one needs to
have many more observations than parameters to be estimated, or the
resulting model will be too noisy to be meaningful (theoretically a
Bayesian model with a very strong prior could be an exception, but
that's such a departure it wouldn't normally be called statistics).
If you were familiar with principal component analysis then you would
know that its application to regression estimation is actually to
reduce the number of parameters to be estimated in a statically
efficient way. But what we are engaged in here is called
*digression*.
> For example, if you have known inputs and a
> single output, you could construct something like
> log speed = a x log power + b x log frontal area +
> c x log headwind. You have say 50 riders and each of them
> has a measured speed, power, area, headwind, and you do some
> kind of multiple regression to find out the coefficients.
>
> If you don't know the inputs but have multiple outputs, you can
> do something like principal component analysis, or the aptitude-test
> factor analysis outlined in the wikipedia page on factor analysis
> that you helpfully pointed out. Here you have several measurements
> for each subject and the goal (at least in PCA) is to reduce the
> dimensionality of the data set. But you need more measurements
> per subject than factors (hidden variables) otherwise the linear
> model is underconstrained.
What you are saying about PCA shows a major misunderstanding of the
topic. PCA in the context of regression modeling nothing more than
the following: 1) do eigenanalysis of matrix of explanatory variables,
2) rotate the equations to the new eigenspace and pick the n
eigenvectors with the largest eigenvalues as your new explanatory
variables (where n is lower than the original number parameters to be
estimated), 3) do statistical fitting as before in the new reduced
subspace, the point being that using eigenvectors insures that all
parameter estimates are now statistically orthogonal and using the ones
with largest eigenvalues insures the least addition of model bias in
the least squares sense.
> The problem with applying this to TTs is that there are only two
> obvious known measurements, speed and distance.
> So you can pile up all the TTs Kloden or Kessler ever did in his career, plot
> speed versus distance, derive a mean relationship, and tell whether
> stage 7 was off the relationship. Or use multiple riders to derive
> the slope of the speed-distance relationship. But this isn't so
> complicated that it needs to be called factor analysis.
There are a lot of different ways of using statistical models to
generate estimated performance ratings. Look at the research article
biblio of this web course for some general ideas:
http://www.stat.sc.edu/~habing/courses/718S05.html
As I noted before, factor analysis doesn't exhaust the range of linear
modeling approaches and linear models are just the simplest type of
statistical approaches to the problem.
Now consider that in order to use the data from cycling efficiently,
one has the further complication of making one of the following
choices:
a) do statistics on a small set of riders and events that feature all
riders in each event, or
b) use some sort of missing data technique for when a give rider misses
and event, or
c) break the results down into pairwise comparisons between riders but
then use GEE (generalized estimating equations) to deal with the
additional covariance resulting from the same individual performances
appearing in differential pairwise comparison data points
> If you want to get really fancy, you can start to introduce a per-stage
> term that attempts to correlate across multiple riders to tell whether
> a given TT was faster or slower than the norm -
I was assuming all along that what we are using as our dependent
variable is either the differential in rider times or the differential
per KM rather than the absolute time. That avoids this particular
issue.
> Since I wasn't the one who guessed that T-Mobile were all a standard
> deviation off an unspecified model, perhaps you should come up with
> a better model.
It depends on my purpose. My purpose in introducing talk of models
into this thread was merely to respond to a query about what sort of
thing I meant when I was speaking about statistical probabilities. It
isn't necessary to pick a specific modeling technique in order to
explain the gist of what I meant.
> BTW,
> 2006 Giro stage 11, 50 km ITT
> 1 Jan Ullrich (Ger) T-Mobile Team 58.48 (51.02 km/h)
> 2 Ivan Basso (Ita) Team CSC 0.28
> 3 Marco Pinotti (Ita) Saunier Duval-Prodir 1.01
> 4 Serguei Gonchar (Ukr) T-Mobile Team 1.09
>
> 2005 Giro stage 8, 45 km
> 1 David Zabriskie (USA) Team CSC 58.31 (46.14 km/h)
> 2 Ivan Basso (Ita) Team CSC 0.17
> 3 Paolo Savoldelli (Ita) Discovery Channel 0.44
> 4 Marzio Bruseghin (Ita) Fassa Bortolo 0.48
> 5 Serguei Gonchar (Ukr) Domina Vacanze 0.51
I guess you are trying to say that Gonchar is known to be a decently
strong rider in TT?? I already knew that, but this doesn't account for
the overall surprise of the T-mobile result.
Howard Kveck
Curtis L. Russell <cur...@md-bicycling.org> wrote:
> Levi's wife hums some stupid song in the morning, it gets stuck in his
> head, he rides like shit. I can see this.
"Knock Three Times?"
--
tanx,
Howard
Never take a tenant with a monkey.
remove YOUR SHOES to reply, ok?
Howard Kveck
Jenko <sally...@gmail.com> wrote:
I wonder which of them was in front on the descent or if they actually split it.
Mark
> In article <ENmdnUxGltbgTSzZ...@comcast.com>,
> Mark <remove.ma...@remove.comcast.this.net> wrote:
>
>
>>benjo maso wrote:
>>
>>>
>>>IMO the point is that the T-Mobile, probably the team with the highest
>>>budget, has a lot of excellent riders. They were expected to help Ullrich,
>>>but now that they are on their own, they are free to race as well as they
>>>can.
>>
>>Good points, but a question: How much of that budget is tied up in
>>Ullrich, and how much of the resources are available to support the
>>remaining riders?
>>
>>Mark
>
>
> Unless you think you can improve a rider by paying him more, what do you
> mean, Mark?
>
> In terms of "rider support resources" for a pro, the one pricey thing
> you can do is wind tunnel time. One way or another, the riders will
> probably be able to get the training and the other resources they need.
>
> Once the race starts, there's not going to be a lot of Ullrich-specific
> resources that can't be redistributed to the new #1. Maybe Jan would
> have a spare TT bike where Klodi might not, but that's about the limit
> of it, I think.
>
> What were you thinking of?
Salary. Maybe that's silly, but to listen to the hype around here,
you'd think Der Jan was getting a big part of T-Mo's budget, in a way
that *can't* be redistributed to the other riders.
Maybe by bringing up budget you meant that T-Mobile had already bought a
bunch of top-drawer riders, and that non-salary support was pretty equal
among teams.
You might be right, but just finishing Bobke's book about tour
conditions in his day, I thought amount of support spending for/during
the tour itself might make a difference.
Mark
h squared
> And he's a career-long TT specialist. I may have this a little wrong,
> but I think he's won five Giro stages, and four of them were TTs.
unless i'm screwed up (possible! :), they were all itts (?)
http://www.cyclingbase.com/palcoureurs.php?id=863&idtitle=1
(note, that website numbers the prologues as "stage 1" so the stage
numbers may be one off from the normal convention, sorry)
1997
*1st 18th stage of the tour of italy, itt (5th final gc)
2nd in the 9th and 21st stages of the tour of spain, itts
2nd in the world championship itt
1998
*1st in the 23rd stage of the tour of italy, itt (10th final gc)
2nd in the 1st and 16th stages of the tour of italy, itts
3rd in the world championship itt
1999
*1st in the 18th stage of the tour of italy, itt (7th final gc)
2nd in the 9th stage of the tour of italy, itt
6th in the world championship itt
2000
4th in the 21st stage of the tour of italy, itt (9th final gc)
4th in the 21st stage of the tour of spain, itt
1st in the world championship itt
2001
4th in the 16th stage of the tour of italy, itt (4th final gc)
2002
3rd in the 10th stage of the tour de france, itt
6th in the 20th stage of the tour de france, itt
2003
*1st in the 21st stage of the tour of italy, itt (8th final gc)
3rd in the 15th stage, itt
2004
*1st in the 14th stage of the tour of italy (2nd final gc)
3rd in the 19th stage of the tour of italy- not an itt
http://www.cyclingnews.com/road/2004/giro04/images/T18_alt.jpg
not trying to nitpick, more that i wanted to post his more memorable
finishes and give him some credit :)
sorry if it's already been done, i CANNOT keep up to date with the posts
around here in july.
h
Ryan Cousineau
Mark <remove.ma...@remove.comcast.this.net> wrote:
I'm doubtful. I think there's a few effects which are correlated but not
causative.
-the salaries for a top team are a major part of the team's annual
budget. You could probably fund a couple of the cheapest ProTour teams
for what Jan collects in salary.
-Similarly, the teams that can afford the Lances, Jans, and Ivans of the
cycling world also tend to surround them with talent.
-A-list teams can do this because of A-list sponsors, and they attract
A-list sponsors because they generate publicity all out of proportion to
the cost of the sponsorship (compared to, say, spending promo dollars on
straight TV ads, for example). Cycling team sponsorship trades exposure
for variance (your team could suck that year, or Jan could get pulled
from the Tour...). AG2R doesn't get Jan-class coverage, and so they get
fewer dollars out of their sponsors.
-the teams that have all that money to spend on top riders and top
domestiques have plenty left over for the ancillary costs of being a pro
team: vehicles, cooks, soigneurs, wind tunnel time, and all those other
little bits.
So a team with a great riders will have a great support system, but a
great support system is not what makes them great. Ivan Basso probably
could have won the Giro d'Italia eating AG2R's food and getting rubbed
down by AG2R's soigneurs, but maybe not with AG2R's riders as his
teammates.
--
Ryan Cousineau rcou...@sfu.ca http://www.wiredcola.com/
"I don't want kids who are thinking about going into mathematics
to think that they have to take drugs to succeed." -Paul Erdos
Ryan Cousineau
Tim Lines <SP...@SPAM.com> wrote:
Do dogs on the Internet dream of electric sheep?
Ryan Cousineau
"Scott" <hendric...@hotmail.com> wrote:
> psycholist wrote:
> > I'm sure you were fishing for this, so I'll oblige you.
> >
> > It's not about the bike. If it were, CSC would have taken all the top
> > spots and Bobby Julich wouldn't have crashed 'cuz everyone knows that
> > Cervelos can do no wrong. At least the trithletes know that over at
> > Slowtwitch.com.
> >
> >
>
> No, that would be Quintana Roo, not Cervelo.
Yeah, well, triathletes are all dumbasses who can't go around corners.
Ryan Cousineau
Stu Fleming <ste...@wic.co.nz> wrote:
> Simon Brooke wrote:
> > in message <1152490093....@s13g2000cwa.googlegroups.com>,
> > Patricio Carlos ('pg...@hotmail.com') wrote:
> >
> >
> >>So are TMO doping more? Or are DSC doping less? Or both? Please help.
> >
> >
> > My /guess/ is that T-Mo are not doping (at least, not illegally). There
> > will be a considerable investigative spotlight on them and on CSC.
> > However, seeing T-Mo's performance has not declined while others have, I
> > think the rather worrying conclusion has to be that T-Mo already were
> > the cleaner team.
> >
>
> Hence my comment about this being like 1991 (or at least, pre 1994). I
> believe (and I may be wrong) that we are seeing a "non-boosting" Tour.
> Maybe a "recovery products only" Tour, but without the ridiculous excess
> of the EPO-era. That's also why I draw the line in the mountains at
> 1740 VAM.
Okay, taking the Millar Line, Maginot Line, or Vayer Line seriously is
the path to madness.
Kunichosis or Laffertism, pick your poison,
Ryan Cousineau
Michael Press <ja...@abc.net> wrote:
Learn something new every day. Those wacky Red Wings fans!
Mind you, after the octopus is cooked, some of us decide it might be
better just to eat the thing,
Canucks fans (scary nickname for you, and one that weirdly parrots the
"Canadiens") don't throw anything on the ice, though a few probably lost
their lunch after watching the team sickeningly lurch out of playoff
contention this year,
b...@mambo.ucolick.org
> b...@mambo.ucolick.org wrote:
> > If you don't know the inputs but have multiple outputs, you can
> > do something like principal component analysis, or the aptitude-test
> > factor analysis outlined in the wikipedia page on factor analysis
> > that you helpfully pointed out. Here you have several measurements
> > for each subject and the goal (at least in PCA) is to reduce the
> > dimensionality of the data set. But you need more measurements
> > per subject than factors (hidden variables) otherwise the linear
> > model is underconstrained.
>
> What you are saying about PCA shows a major misunderstanding of the
> topic. PCA in the context of regression modeling nothing more than
> the following: 1) do eigenanalysis of matrix of explanatory variables,
> 2) rotate the equations to the new eigenspace and pick the n
> eigenvectors with the largest eigenvalues as your new explanatory
> variables (where n is lower than the original number parameters to be
> estimated), 3) do statistical fitting as before in the new reduced
> subspace, the point being that using eigenvectors insures that all
> parameter estimates are now statistically orthogonal and using the ones
> with largest eigenvalues insures the least addition of model bias in
> the least squares sense.
Maybe you thought I was saying something more profound
about PCA than I was? Here's a couple of examples.
Suppose each "subject" is a collection of a few hundred
or thousand measurements, such as rainfall in a given
month at several hundred positions on the globe, or
light intensity at several thousand wavelengths in the
spectrum of a galaxy:
http://trmm.jpl.nasa.gov/global/
http://arxiv.org/abs/astro-ph/0305587
You have a bunch of these collections, from different
months, or different galaxies. Each collection is a
vector in a few hundred or thousand dimensional space.
If you do a PCA on the ensemble of vectors, you can (in
either of these problems) find a small number of
eigenvectors that account for most of the variance, so
each vector can be approximated by a linear
combination of some small number n eigenvectors
(similar to what you described), rather than hundreds
or thousands of numbers. This is what I meant by
"reduce the dimensionality." One of the utilities of
this is that the small number of important eigenvectors
often have a physical meaning which helps you
understand the problem, if you set it up right.
Not that I'm claiming to know how to set this up
for analyzing time trials.
But as you said, we are digressing, so I'll
shut up now.
Donald Munro
> Do dogs on the Internet dream of electric sheep?
More importantly can virtual dogs be used as a doping excuse ?
Tuschinski
> Tour 2004
Actually, it's not unseen for the team of the Winner to be on top of
the heap in a TT
I saw it with:
Renault-Gitane
Banesto
T-Mobile
USPostal
This would also imply that T-Mobile will win this tour^^
To elaborate a bit more, if Basso had been there noone would be
surpirsed if those positions had been taken by CSC
(Vandevelde,Zabriskie, Julich, Sastre, Voigt and Basso are all decent
TT riders). And tbh, look at Disco, they are all good TTers as well, I
wouldn't be shocked had they come up with a similar performance. This
time it was T-Mobile, wonder how the mountains will go.
Donald Munro
> Okay, taking the Millar Line, Maginot Line, or Vayer Line seriously is
> the path to madness.
>
> Kunichosis or Laffertism, pick your poison,
Better ask Hobson to generate a few more choices.
Ryan Cousineau
Donald Munro <fat-d...@hotmail.com> wrote:
No, but if your virtual dog dies it's an indication of gene doping.