Power Data Estimates for the climbing stages

[zigmeister]

i dont have to own up. horner has to.

the testers showed up in his team hotel where all his teammates were but horner.

...read tyler on how they manipulated the whereabouts system by entering changes in the latest possible moment to confuse the testers.

Looking through the thread, don't think this was addressed, but was publicly of course. His whereabouts were properly provided to the appropriate people. The testers never got the update of his new hotel where he was with his wife, so they showed up at the wrong place.

He was where he said he was, they screwed up on the testing side, which is no surprise based on their sketchy and questionable past and procedures.

Anywho...nice try.

 

[acoggan]

Study reveals the scope of imprecision in climbing power estimates based on ascension rates:

http://journals.humankinetics.com/A...umentItem/Millet_IJSPP_2013_0320-in press.pdf

From that paper - a chart plotting the percentage difference in power estimates versus actual power measured with SRM power meter:

Main reason for the imprecision? Inability to quantify impact of variable air drag. Gee, who'd have thunk it?

Wind is thine enemy.

 

[Alex Simmons/RST]

Notice there is a bigger variation for the wheelsuckers in the 250-350W range than the winners/leaders in the 350W+ range. Its not super important to estimate pack fodders outputs because nobody cares what power they put out.

Yet people use pack fodder who release power meter data as a means to validate power estimates for the leaders.

Conclusions: Aerodynamic drag (affected by wind velocity and orientation, frontal area, drafting and speed) is the most confounding factor. The mean estimated values are close to the PO values measured by powermeters, but the random error is between ±6% and ±10%.

Moreover, at the POs (>400 W) produced by professional riders, this error is likely to be higher. This observation calls into question the validity of releasing individual values without the reporting of the range of random errors.

 

[Merckx index]

Moreover, at the POs (>400 W) produced by professional riders, this error is likely to be higher. This observation calls into question the validity of releasing individual values without the reporting of the range of random errors.

What do they mean by "releasing"? Anyone can estimate power values by timing the climb. No official release of these values is necessary.

So main findings:

1) The mean difference was a little less than 1%.
2) Less than 5% of the individual differences were >10%, and (by my rough count), less than 20% of the differences were more than a negative 5%--i.e., only in this small minority did the estimate exceed the SRM value by more than 5%. This figure drops to about 2% in low winds.

My conclusions:

1) Since the mean difference is quite small, comparison of several well-documented climbs in different years, referring to multiple climbers, should provide a good indication of changes in power over various time periods, with due caution for other factors that may be different in these years.
2) In light winds, power estimates are very likely to be good enough to support any conclusions that would be implied with SRM data. Even in strong winds, the differences are usually going to be within, say, 0.30 watts/kg.
3) To the extent that these data do suggest caution in interpreting power estimates, though, they underscore the disingenuousness of certain teams that claim to be clean but will still not release SRM data in key time periods. You can't have it both ways. If power estimates are not to be trusted, than any failure to release SRM data is akin to someone suspected of doping refusing to be tested.

 

[Alex Simmons/RST]

If power estimates are not to be trusted, than any failure to release SRM data is akin to someone suspected of doping refusing to be tested.

Why?

Since power data will never be a basis for proof of doping under any anti-doping code, then this makes no sense.

Proof of doping will always be a haematological matter as well as one for accepted investigative processes (e.g. admissions and other non-dope testing means of establishing doping).

Improving the accuracy of our knowledge of riders' power outputs doesn't improve our ability to detect dopers, and neither will power data ever result in a doping infraction.

And knowing power accurately can never rule out doping.

So what's the point in anyone releasing it other than to present a meaningless veil of anti-doping transparency to feed the twitter minions?


So according to you, if I (say) don't release my power data, you then presume I am guilty of hiding some form of doping activity?

If I (say) do release it, then every wannabe analyst will form their biased opinion in any case.

Will we any better informed than before?
No.

Will any doping infractions result from release of data?
No.

Do we already know who the likely targets for extra scrutiny are?
Of course we do.

 

[Ferminal]

Good stuff. Would be nice if they release the raw data when it's published.

 

[Merckx index]

Why?

Since power data will never be a basis for proof of doping under any anti-doping code, then this makes no sense.

Proof of doping will always be a haematological matter as well as one for accepted investigative processes (e.g. admissions and other non-dope testing means of establishing doping).

Improving the accuracy of our knowledge of riders' power outputs doesn't improve our ability to detect dopers, and neither will power data ever result in a doping infraction.

And knowing power accurately can never rule out doping.

So what's the point in anyone releasing it other than to present a meaningless veil of anti-doping transparency to feed the twitter minions?

I said "akin to". Power data are about all we have other than tests which are generally easy to beat. If you have seen even a few of my posts on this subject, you know that I'm not recommending power data be used as a basis of a sanction. But of course they will be used by fans to form opinions.

In addition, some people highly-placed in the sport, like JV, use power data to argue that the peloton is cleaner. In fact, I don't know anyone who argues that the peloton is cleaner who doesn't use power data as the main argument, do you? Walsh, for example, while arguing that the data are too imperfect to conclude Froome is doping, has no problem using the same data to conclude the peloton is cleaner, and even that Froome's climbs are evidence of being clean.

What other argument could be used? The number of positives? If it's lower, does that mean the peloton is cleaner, or just that the riders are better at beating the tests? Anyone who thinks that power data are useless must also be totally agnostic on the question of whether the peloton is cleaner. Maybe you are, but very few other people following the sport are.

So according to you, if I (say) don't release my power data, you then presume I am guilty of hiding some form of doping activity?

If I (say) do release it, then every wannabe analyst will form their biased opinion in any case.

Your false assumption here is that if we don't release the power data, followers of the sport will not form any opinion at all. Opinions are always going to be out there. If you think that by not releasing data, opinions will not be formed, you must have missed the Sky/Froome threads entirely.

Again, if power data are meaningless, why did Sky release data for Froome from 2011-2013? What was the point? Do you think they thought the data were totally worthless, and they released them just to placate a handful of critics in the Clinic? Even if that were the entirely the case, that shows that very strong opinions were formed even before any data were released. I don't see how those data could possibly make the situation worse. On the contrary, if SRM data are so much more reliable than timing climbs for individual riders, then why would you not want to release them? Do you think if you give better data to people following the sport, their opinions will become even more misinformed?

Andy Coggan at least makes a debatable argument against releasing SRM data--that it's too expensive. That may or may not be a dealbreaker, but that is very different from saying the data are useless.

Will we any better informed than before?
No.

Of course we will. As the article concludes, under low wind conditions, as well as in analysis of a large pool of riders, the differences between SRM and estimates from time are reduced to the point where we can make reasonable conclusions. Again, I'm not saying this will allow sanctions, but it will make us better informed--particularly about the overall state of the peloton.

I will grant you that there will be plenty of misinterpretation of any data. But so what? The fans in effect pay the salaries of the riders. Don't they have the right to see data bearing on how clean they are?

Do we already know who the likely targets for extra scrutiny are?
Of course we do.

Not always. I disagree very strongly with the implication that data like these can't help target riders more intelligently. It isn't always about the elite riders, it's also about sudden major improvements. People were talking about DiLuca in the Giro even before he was popped.

 

[Dear Wiggo]

Proof of doping will always be a haematological matter as well as one for accepted investigative processes (e.g. admissions and other non-dope testing means of establishing doping).

Haematological or urinary.

 

[Alex Simmons/RST]

Haematological or urinary.

True, but you get my drift.

 

[Alex Simmons/RST]

I said "akin to". Power data are about all we have other than tests which are generally easy to beat.

Then the issue is sorting out testing process and effectiveness of investigative measures rather than looking to power data that doesn't really solve the problem, no matter how accurate it might be.

If you have seen even a few of my posts on this subject, you know that I'm not recommending power data be used as a basis of a sanction. But of course they will be used by fans to form opinions.

Well I think fans' opinions should more be based on relevant evidence rather than idle speculation. But that is, of course, just my opinion

In addition, some people highly-placed in the sport, like JV, use power data to argue that the peloton is cleaner. In fact, I don't know anyone who argues that the peloton is cleaner who doesn't use power data as the main argument, do you? Walsh, for example, while arguing that the data are too imperfect to conclude Froome is doping, has no problem using the same data to conclude the peloton is cleaner, and even that Froome's climbs are evidence of being clean.

There is a very large difference between using ascension rates (and W/kg estimates) for population level assessments of the changes in physiological impact of doping over the last 30 years, to undertaking individual rider assessments.

As I have posted before, the climbing speeds have gone up and down over the years, and this corresponds closely with the level of EPO use and other significantly effective forms of blood doping.

But that tells us nothing about the level of doping, only the level of physiological impact doping has had on the professional peloton as a whole. All we can reasonably say is that the performance boost being attained has reduced, not that doping is more or less prevalent.

So on that basis, I totally agree, such statement by high profile people in the sport are somewhat falacious.

As for individual rider assessments, the problem is that the rate of performance change from current doping practices isn't significantly different to what could plausibly be argued as being attainable from non-doping means. I mean my racing W/kg would vary 15+% through a season. Does that make me suspicious?

What other argument could be used? The number of positives? If it's lower, does that mean the peloton is cleaner, or just that the riders are better at beating the tests? Anyone who thinks that power data are useless must also be totally agnostic on the question of whether the peloton is cleaner. Maybe you are, but very few other people following the sport are.

I think I've covered that above. IOW it doesn't deal with these issues either and it's a mistake to think it does.

Your false assumption here is that if we don't release the power data, followers of the sport will not form any opinion at all. Opinions are always going to be out there. If you think that by not releasing data, opinions will not be formed, you must have missed the Sky/Froome threads entirely.

Again, if power data are meaningless, why did Sky release data for Froome from 2011-2013? What was the point? Do you think they thought the data were totally worthless, and they released them just to placate a handful of critics in the Clinic? Even if that were the entirely the case, that shows that very strong opinions were formed even before any data were released. I don't see how those data could possibly make the situation worse. On the contrary, if SRM data are so much more reliable than timing climbs for individual riders, then why would you not want to release them? Do you think if you give better data to people following the sport, their opinions will become even more misinformed?

I agree that opinions will be formed whether or not such data are released, but I don't think the data will make those opinions any more or less valid. For starters we have people who proclaim there is a definitive W/kg mark = doping, which is of course just opinion, when there really isn't any such clear line.

As for why Sky did what they did, I have no idea. Media/PR reasons presumably. It doesn't pass the plausibility test for proving (sans-)doping status either way. I suppose damned if they do and damned if they don't. It's a no win strategy, because it doesn't prove anything.

Andy Coggan at least makes a debatable argument against releasing SRM data--that it's too expensive. That may or may not be a dealbreaker, but that is very different from saying the data are useless.

He may have, I don't specifically recall, although I definitely made this comment in my The Elusive Dopeometer blog post.

Of course we will. As the article concludes, under low wind conditions, as well as in analysis of a large pool of riders, the differences between SRM and estimates from time are reduced to the point where we can make reasonable conclusions. Again, I'm not saying this will allow sanctions, but it will make us better informed--particularly about the overall state of the peloton.

Yes, we can make sensible conclusions about a population of riders, and as you rightly point out, this is inadequate for individual sanction.

A key point is knowing when such models are valid. As is evidenced by this thread, most don't get it and is why I have regularly suggested people put estimated error ranges on their W/kg numbers. The less well know the conditions, the greater the error.

I will grant you that there will be plenty of misinterpretation of any data. But so what? The fans in effect pay the salaries of the riders. Don't they have the right to see data bearing on how clean they are?

They have the right to know if riders are clean, but power data doesn't tell us that, which is my point.

All we can really say is that the impact of doping on physiological performance has declined, not that doping has declined.

Not always. I disagree very strongly with the implication that data like these can't help target riders more intelligently. It isn't always about the elite riders, it's also about sudden major improvements. People were talking about DiLuca in the Giro even before he was popped.

Do we need power meter data to see such changes in performance? Aren't race results sufficient?

Is knowing who to target really that hard to figure out?

Did any of those sufficiently prescient about Di Luca need his SRM data to suggest he might need extra scrutiny than a pro might already receive?


I don't have a solution or suggestion to what's needed to prove sans-doping status, because it's an unobtainable outcome. It's a dilemma, I agree.

 

[zigmeister]

I'm still waiting for this thread to provide actual proof of doping. A bunch of power estimates with using "math" and pseudo science doesn't refer to reality.

 

[JMBeaushrimp]

I'm still waiting for this thread to provide actual proof of doping. A bunch of power estimates with using "math" and pseudo science doesn't refer to reality.

Yeah. Stupid "math" never proved anything...

I may be wrong, but it's almost sounding semantic. Are MI and Simmons arguing the ends against the middle? I'm not sure. I am sure, though, that both of their arguments should lend some increased credibility and accuracy to assessing performances from my vantage point - that being watching from the side of the road, or on my sofa...

Thanks to both.

 

[Alex Simmons/RST]

Yeah. Stupid "math" never proved anything...

I may be wrong, but it's almost sounding semantic. Are MI and Simmons arguing the ends against the middle? I'm not sure. I am sure, though, that both of their arguments should lend some increased credibility and accuracy to assessing performances from my vantage point - that being watching from the side of the road, or on my sofa...

Thanks to both.

I imagine we agree on far more than we disagree, but it's points of difference that are highlighted/focussed on as that is the nature of interweb discussions.

Still, this place is after all just a virtual pub, and not a particularly sound forum for resolving the issue. Pub chats can be fun or annoying.

 

[Le breton]

Study reveals the scope of imprecision in climbing power estimates based on ascension rates:

http://journals.humankinetics.com/A...umentItem/Millet_IJSPP_2013_0320-in press.pdf

From that paper - a chart plotting the percentage difference in power estimates versus actual power measured with SRM power meter:

Main reason for the imprecision? Inability to quantify impact of variable air drag. Gee, who'd have thunk it?

I see that Grappe is one of the authors, which I consider to be a rather ominous sign.
(he is the man who in L'Equipe sang to the tune of his master L.A. during the 2001 TdF and he has always struck me as being seriously physics-challenged )
Apart from that he maybe a nicer guy than his arch-enemy A. Vayer.

You will also have noticed that the article has not been proof read, and presumably not peer-reviewd either.

Still, I'll read it when I have time to waste. Between Christmas and New year maybe, obviously not a priority considering the names of the authors.
Added next day :I stand corrected
Apparently the article was peer-reviewed
The article appears here in its accepted, peer-reviewed form, as it was
provided by the submitting author. It has not been copyedited,
proofread, or formatted by the publisher.

 

[Ferminal]

Vayer seems to be going ape**** about something.

 

[Ferminal]

http://veloclinic.tumblr.com/post/69766365103/in-one-of-the-odder-twists-of-late-it-turns-out

lol, maybe Vayer can release the data then.

 

[Fearless Greg Lemond]

So, it comes down to this:

versus

??

Where is Copernicus when you need him. Oh yeah, burned at the stake.

 

[Le breton]

http://veloclinic.tumblr.com/post/69766365103/in-one-of-the-odder-twists-of-late-it-turns-out

lol, maybe Vayer can release the data then.

Many thanks for pointing out this veloclinic article.
So, this humankinetics article is just the latest skirmish in the Grappe/Millet battle against Vayer?

I can't spend hours trying to unravel all this.
Since I don't have the time it comes down to a matter of trust.
On one side :
I trust Portoleau, as for Vayer I know that he overstates his case quite too often.
On the other side :
I distrust Grappe 1) because he was totally blind (or corrupted by L.A.) in 2001 and 2) does not seem to know/understand physics (from tidbits I picked up here and there), therefore I suspect he just uses recipes from handbooks to get his results.

As for my almost neighbour Millet, I wonder if there are two of them, one in Lausanne and one in Montpellier, or if it is the one and same Grégoire.
The one in Montpellier had Vayer as a co-author on some paper about 10-years ago (and acknowledged my help, but I don't know how exactly my email had helped him )

 

[Ferminal]

Many thanks for pointing out this veloclinic article.
So, this humankinetics article is just the latest skirmish in the Grappe/Millet battle against Vayer?

I can't spend hours trying to unravel all this.
Since I don't have the time it comes down to a matter of trust.
On one side :
I trust Portoleau, as for Vayer I know that he overstates his case quite too often.
On the other side :
I distrust Grappe 1) because he was totally blind (or corrupted by L.A.) in 2001 and 2) does not seem to know/understand physics (from tidbits I picked up here and there), therefore I suspect he just uses recipes from handbooks to get his results.

As for my almost neighbour Millet, I wonder if there are two of them, one in Lausanne and one in Montpellier, or if it is the one and same Grégoire.
The one in Montpellier had Vayer as a co-author on some paper about 10-years ago (and acknowledged my help, but I don't know how exactly my email had helped him )

Well I look forward to your insight if you ever find the time to do more unraveling! I agree that Vayer and Grappe's views should be taken with caution (not to say that their data is invalid).

Interesting that Julien Pinot, Thibaut's brother, is a PhD student under Grappe. Both work as coaches for FDJ.

 

[Gregga]

The clash goes on on twitter between Grappe, Millet and Vayer.
https://twitter.com/veloclinic/status/411186585267220480

I still don't get why these french guys want so much Vayer to shut up

 

[Le breton]

The clash goes on on twitter between Grappe, Millet and Vayer.
https://twitter.com/veloclinic/status/411186585267220480

I still don't get why these french guys want so much Vayer to shut up

This is funny.
It looks like they are talking about the study done many years for which I e-mailed some comments to Millet. Wish I could retrieve that e-mail, it's on a corrupted USB stick. Anyway, they didn't have that much data then, guess they were just beginning their study.

As for the conflict, it seems that Vayer is not the nicest of fellows in everyday life, apparently from what I've read, he is quite arrogant. But I would love to get my hands on his stash of stress tests at Festina, that would give us the real dope on the effects of EPO + this + that

 

[thehog]

This is funny.
It looks like they are talking about the study done many years for which I e-mailed some comments to Millet. Wish I could retrieve that e-mail, it's on a corrupted USB stick. Anyway, they didn't have that much data then, guess they were just beginning their study.

As for the conflict, it seems that Vayer is not the nicest of fellows in everyday life, apparently from what I've read, he is quite arrogant. But I would love to get my hands on his stash of stress tests at Festina, that would give us the real dope on the effects of EPO + this + that

Well he's taking the p1ss out of the fact that Grappe adjusted the figures to make Armstrong look better.

Grappe whom now is the man of the moment.

 

[Fearless Greg Lemond]

But I would love to get my hands on his stash of stress tests at Festina, that would give us the real dope on the effects of EPO + this + that

+ 200, but Festina always had headwinds.

 

[thehog]

+ 200, but Festina always had headwinds.

How long before SRM build a wind meter into their units?

Athletics had 40 years of data with their meters.

These days they are small enough to use in a head unit or probably better on the cadence sensor to get bi-directional measurements.

Better go register my startup

 

[BroDeal]

I skimmed through the paper. What I got from it is that estimates in little wind can be quite accurate. I did not find the conclusions matched some of what was in the paper. The word "dishonest" was used to refer to people using such estimates. I found this a bit strange. The apparent bad blood between Vayer and the authors make the use seem like an attack on a rival. They also seem to shy away from the obvious, that error could be made very small by using multiple climbs.

The bottom line is whether you want to go through convoluted math to estimate Wattages or not, when a rider is repeatedly doing times on equivalent to the best dopers of the blood vector doping era then something is wrong.