The Froome Files, test data only thread

[acoggan]

Thanks for the reply: you arguably are the most qualified on this forum wrt sports science. But here we go again, and I made a big fuss about it back in July: what is the mystery? Why are we made to believe that Sky/Froome don't know the exact numbers? One would think that at this level, athletes would be constantly monitored.

That would be a mistaken belief. Physiological testing costs time and money, and only provides information of limited usefulness to coaches and athletes (vs. scientists). IMO it is therefore quite understandable why pro cycling teams don't invest heavily (if at all) in such endeavors, especially when you consider that 1) their athletes (employees) often live great distances from any sort of headquarters, 2) compared to, e.g., American football, F1, etc., large sums of money aren't available, and 3) equally viable alternatives (i.e., measuring power directly) exist.

One way to look at it is this: if pro cycling teams actually saw significant value in physiological testing and hence devoted lots of resources to it (or, e.g., wind tunnel testing), then why do all the world's experts in such areas (e.g., Dave Martin, John Cobb) seem to work elsewhere?

 

[harryh]

Really refresing twitter comments by @JeroenSwart today

 

[gillan1969]

Re: Re:

Bold formatting fail. Why would there be two *almost* identical versions of this "fax"? Ropey as fvck.

Binder ring holes in two different positions

Two faxes, one bold, one not, odd shaped top page, binder holes misaligned.

the dodgy dossier part deux first class....

 

[Ventoux Boar]

So, he was never a donkey, or the most transformed cyclist in history. And it seems he succeeded in spite of Brailsford and Kerrison's world class management and sports science: an early vuelta crash and he's out the door after all. Looks like the theory that Tim and Dave are way overrated just got stronger. The only way to succeed at Sky is to have freakish physiology. This explains much.

 

[Tonton]

Re: Re:

Froome's VO2Max was tested at 84, not 88. No ifs, no buts. It's 84.

That's an overly-simplistic point-of-view. Froome's VO2max was measured in August as being 5.91 L/min. What it was when he won the Tour in July will obviously depend upon not only any changes in his cardiovascular fitness between the two months, but also any changes in his body mass. It is therefore quite reasonable to ask the question, "what would his VO2max have been in July at his reported body mass of 67 kg, assuming no changes in the absolute value?", with the answer to that question being "88 mL/min/kg."

(Note that another quite reasonable question to ask - and one that I'm a bit surprised wasn't addressed in the GSK report - is "how does his VO2max in mL/min/kg lean body mass compare between 2015 and 2007?" But, as emphasized in the data release it's hard to directly compare the two datasets due to possible differences in how the data were collected, and they may be saving that calculation for the final paper, either due to time or just to keep things simpler for non-scientists to understand.)

I get the idea of a formula that would take lesser weight into account and therefore raise VO2Max, yet it seems that other values may be affected, i.e. efficiency, in a negative way. I think that's what you alluded to with "assuming no changes".

I'm concerned that a recalculation like this results in a mere indication of potential (as in "best case scenario"), rather than actual performance (measured). Many articles following the release of data mentioned LeMond, but was he in optimum shape when recording his 92? Should that number be adjusted too? And now we're not talking about measured performance, but fantasy cycling. Some have done similar manipulations to estimate Carl Lewis' 100m potential by putting together his best 10m splits. The answer is 9"86: what he did.

What you score is what you are, IMO. The rest is conjecture. I don't like the idea of manipulating numbers to fit an agenda.

 

[djpbaltimore]

Thanks for the tip, harryh. Interesting, but not surprising, how some people are so quick to attack the messenger. Now Swart is seen as an 'independent' tester.

 

[gillan1969]

Jeroen Swart ‏@JeroenSwart 5 hrs5 hours ago

.@acoggan having never tested an athlete of @chrisfroome caliber we mistakenly set the submaximal test length just a little too short


One for the scientists....whats the implications of the test lenght being too short (if any)? and how did/could they make this mistake i.e. is it an easy mistake to make as it sound quite basic

....or should I await the peer reviewed paper...

 

[gillan1969]

Re: Re:

Bold formatting fail. Why would there be two *almost* identical versions of this "fax"? Ropey as fvck.

Binder ring holes in two different positions

Two faxes, one bold, one not, odd shaped top page, binder holes misaligned.

the dodgy dossier part deux first class....

you wait 8 years for your test results then two come along at once.....sorry couldn't resist

 

[thehog]

Jeroen Swart ‏@JeroenSwart 5 hrs5 hours ago

.@acoggan having never tested an athlete of @chrisfroome caliber we mistakenly set the submaximal test length just a little too short


One for the scientists....whats the implications of the test lenght being too short (if any)? and how did/could they make this mistake i.e. is it an easy mistake to make as it sound quite basic

....or should I await the peer reviewed paper...

Truth be told, testing is generally conducted over several sessions. This was a one off test on one day thus it's open to mishaps, testing anomalies etc. It happens. As long as the output data is not fundamentally disturbed then it remains valid.

All testing should withstand scrutiny, in this case there are some question over the method. Cutting the test slightly short doesn't discount the entire effort.

I don't have a massive issue with the the 2015 test, I just don't believe the 2007 data to be valid. It's just too odd along with those faxes looking awfully suspect.

Just my opinion.

 

[djpbaltimore]

I agree with thehog that the 2007 data set remains the most questionable piece of the puzzle. But, I will wait until the paper comes out to see what is discussed in that regard before making a judgment on its relative merit.

 

[thehog]

Re:

I agree with thehog that the 2007 data set remains the most questionable piece of the puzzle. But, I will wait until the paper comes out to see what is discussed in that regard before making a judgment on its relative merit.

I don't have a problem with Swart either. I think he conducted his testing well. It is his moment in the sun and he is getting some notoriety for it. He is human like all of us. I've seen him on Twitter, he is slightly swaying into making conclusions on doping which he can't really do. He can only speak to the tests and output he conducted. Vayer as well.

Swart didn't conduct the 2007 tests but he might have seen a lot more data we have by way of fax, therefore some of his conclusions might based on sound logic.

However there are too many poor performances from Froome in between (2007-2011) which bring the 2007 data into question. Swart shouldn't ignore that. Race data is data then same as lab test data, just in a different form.

At this point; if he is solely working off that one scuffed up/photoshopped sheet that we see then he's not in a position to make a conclusion. He can only really say that the physiology from the 2007 test matches the 2015 output. He really should put an NB on the 2007 data though.

 

[harryh]

Jeroen Swart ‏@JeroenSwart 5 hrs5 hours ago

.@acoggan having never tested an athlete of @chrisfroome caliber we mistakenly set the submaximal test length just a little too short


One for the scientists....whats the implications of the test lenght being too short (if any)? and how did/could they make this mistake i.e. is it an easy mistake to make as it sound quite basic

....or should I await the peer reviewed paper...

 

[gillan1969]

Jeroen Swart ‏@JeroenSwart 5 hrs5 hours ago

.@acoggan having never tested an athlete of @chrisfroome caliber we mistakenly set the submaximal test length just a little too short


One for the scientists....whats the implications of the test lenght being too short (if any)? and how did/could they make this mistake i.e. is it an easy mistake to make as it sound quite basic

....or should I await the peer reviewed paper...

Truth be told, testing is conducted over several sessions. This was a one odd test on one day thus it's open to mishaps, testing anomalies etc. It happens. As long as the output data is not fundamentally disturbed then it remains valid.

All testing should withstand scrutiny, in this case there are some question over the method. Cutting the test slightly short doesn't discount the entire effort.

I don't have a massive issue with the the 2015 test, I just don't believe the 2007 data to be valid. It's just too odd along with those faxes looking awfully suspect.

Just my opinion.

no....fair points...another one here


Jeroen Swart
‏@JeroenSwart

.@SamueleMarcora @BStulberg @Scienceofsport lets all be honest please. The call was for Chris to undergo physiological testing. Which he has

is this not the scientist being used by the 'client' and hence de facto losing objectivity ....That wasn't really the call, or it was only insofar as a poltician ask themselves the question they want to answer (or is the only answer they have)..the real call was 'please let us see somthing which explains how he 'transformed''. He didn't need to do test, the results of which everyone knew beforehand i.e. a GT winner...Now, in news management terms, we have independent scientist quite rightly defending his work against people who wanted something else...this plays into the sky narrative of psuedo-scientists, sour grapes and haters.....

Swart being played for a sucker methinks...

 

[thehog]

Jeroen Swart ‏@JeroenSwart 5 hrs5 hours ago

.@acoggan having never tested an athlete of @chrisfroome caliber we mistakenly set the submaximal test length just a little too short


One for the scientists....whats the implications of the test lenght being too short (if any)? and how did/could they make this mistake i.e. is it an easy mistake to make as it sound quite basic

....or should I await the peer reviewed paper...

Truth be told, testing is conducted over several sessions. This was a one odd test on one day thus it's open to mishaps, testing anomalies etc. It happens. As long as the output data is not fundamentally disturbed then it remains valid.

All testing should withstand scrutiny, in this case there are some question over the method. Cutting the test slightly short doesn't discount the entire effort.

I don't have a massive issue with the the 2015 test, I just don't believe the 2007 data to be valid. It's just too odd along with those faxes looking awfully suspect.

Just my opinion.

no....fair points...another one here


Jeroen Swart
‏@JeroenSwart

.@SamueleMarcora @BStulberg @Scienceofsport lets all be honest please. The call was for Chris to undergo physiological testing. Which he has

is this not the scientist being used by the 'client' and hence de facto losing objectivity ....That wasn't really the call, or it was only insofar as a poltician ask themselves the question they want to answer (or is the only answer they have)..the real call was 'please let us see somthing which explains how he 'transformed''. He didn't need to do test, the results of which everyone knew beforehand i.e. a GT winner...Now, in news management terms, we have independent scientist quite rightly defending his work against people who wanted something else...this plays into the sky narrative of psuedo-scientists, sour grapes and haters.....

Swart being played for a sucker methinks...

Swart freely admits Froome is his biggest client yet. Maybe this will be a boon in business for the GSK centre. They are a "for profit" outfit.

The one issue I see with him is he's been thrust into the limelight, he's now in Esquire etc. That's a big deal for a scientist. I hope he keeps his head.

I looked through the GSK PDF and did note they were using a Turbo trainer with ComputerTrainer power device on the rear wheel. Normally you'd use a "rig" setup so the calibration and regulation of the device itself can be measured and monitored for error rates.

For home use, this is fine but for a serious test is a little amateurish.

 

[acoggan]

Jeroen Swart ‏@JeroenSwart 5 hrs5 hours ago

.@acoggan having never tested an athlete of @chrisfroome caliber we mistakenly set the submaximal test length just a little too short


One for the scientists....whats the implications of the test lenght being too short (if any)? and how did/could they make this mistake i.e. is it an easy mistake to make as it sound quite basic

....or should I await the peer reviewed paper...

The only issue is that they shouldn't have applied the Dmax/modified Dmax methods to calculate lactate threshold, since those approaches are predicated on continuing the test all the way to fatigue. IOW, my comment was about how the data were analyzed, not the actual numbers themselves. Given, however, that 1) the lactate data were analyzed multiple ways, 2) Froome did squeak over OBLA, and 3) the raw data are presented, it's really a minor point.

IOW, it caught my attention simply because you don't expect silly little mistakes like this to occur, not anything to do with Froome/Froome's physiology.

For the final paper, they should just ditch the Dmax analyses, as they really don't add anything to the "story" but do open them up to (again, minor) criticism.

EDIT: This post describes/illustrates the Dmax method, which will hopefully help make the issue more clear:

http://www.trainingandracingwithapowermeter.com/2010/08/estimation-of-functional-threshold.html

 

[acoggan]

Re: Re:

(Note that another quite reasonable question to ask - and one that I'm a bit surprised wasn't addressed in the GSK report - is "how does his VO2max in mL/min/kg lean body mass compare between 2015 and 2007?"

Answering my own question:

2007 2015
VO2max (L/min) 6.06 5.91
Body mass (kg) 75.6 69.9
% body fat 16.9 9.8
Lean body mass (kg) 62.8 63.0
VO2max (mL/min/kg LBM) 96.5 93.7

 

[harryh]

I looked through the GSK PDF and did note they were using a Turbo trainer with ComputerTrainer power device on the rear wheel. Normally you'd use a "rig" setup so the calibration and regulation of the device itself can be measured and monitored for error rates.

For home use, this is fine but for a serious test is a little amateurish.

nope

https://twitter.com/fredgrappe/status/672716021191561216?lang=fr

 

[acoggan]

I looked through the GSK PDF and did note they were using a Turbo trainer with ComputerTrainer power device on the rear wheel. Normally you'd use a "rig" setup so the calibration and regulation of the device itself can be measured and monitored for error rates.

For home use, this is fine but for a serious test is a little amateurish.

nope

https://twitter.com/fredgrappe/status/672716021191561216?lang=fr

Yeah, no biggie using the CT, at least given that they had crank power data as well. (Anybody know if Froome used non-round rings during the test? I can't tell from the pictures.) I definitely wouldn't trust the CT data w/o that confirmation, however, as their calibration can vary widely.

 

[thehog]

I looked through the GSK PDF and did note they were using a Turbo trainer with ComputerTrainer power device on the rear wheel. Normally you'd use a "rig" setup so the calibration and regulation of the device itself can be measured and monitored for error rates.

For home use, this is fine but for a serious test is a little amateurish.

nope

https://twitter.com/fredgrappe/status/672716021191561216?lang=fr

Yeah, no biggie using the CT, at least given that they had crank power data as well. (Anybody know if Froome used non-round rings during the test? I can't tell from the pictures.) I definitely wouldn't trust the CT data w/o that confirmation, however, as their calibration can vary widely.

Agreed I don't have a huge issue with it, just commenting on the setup. A lab generally would have a rig device so like you mention the setup/components of subjects bike does not come into question or input skew.

Per the rings; I can't really tell...

Sadly no information in the deck:

 

[robin440]

Definitely Osymetric chainrings on that picture!

 

[Irondan]

Re:

Definitely Osymetric chainrings on that picture!

I disagree, the chain rings look around but it's very difficult to tell for certain.

 

[thehog]

Re: Re:

Definitely Osymetric chainrings on that picture!

I disagree, the chain rings look around but it's very difficult to tell for certain.

Froome using Osymetric rings:

Video of test: https://youtu.be/JzNIH3LSDMM

 

[acoggan]

Re: Re:

Definitely Osymetric chainrings on that picture!

I disagree, the chain rings look around but it's very difficult to tell for certain.

Yes he is using Osymetric rings:

Video of test: https://youtu.be/JzNIH3LSDMM

Interesting. So between the non-round rings resulting in overestimation of power and the fact that the cranks are "upstream" in the drivetrain, you'd expect those numbers to be somewhat higher than that reported by the CT. They were, however, nearly identical, which makes me doubt the absolute accuracy of either set of numbers. OTOH, they definitely seem to be in the right ballpark, so again not that big of a deal (after all, the point of the testing was to provide physiological, not performance, data).

 

[thehog]

Re: Re:

Definitely Osymetric chainrings on that picture!

I disagree, the chain rings look around but it's very difficult to tell for certain.

Yes he is using Osymetric rings:

Video of test: https://youtu.be/JzNIH3LSDMM

Interesting. So between the non-round rings resulting in overestimation of power and the fact that the cranks are "upstream" in the drivetrain, you'd expect those numbers to be somewhat higher than that reported by the CT. They were, however, nearly identical, which makes me doubt the absolute accuracy of either set of numbers. OTOH, they definitely seem to be in the right ballpark, so again not that big of a deal (after all, the point of the testing was to provide physiological, not performance, data).

True, I guess based on the Sky quoted "6% overestimation" variance, Froome was stronger in 2007 than 2015

I jest, hence why they should have used a rig, or picked out stock components. It's not like he needed 6 hours of value from the rings.

 

[armchairclimber]

At least with the set up they used, it's similar to those Froome uses on the road.