power as dopeometer

[Ferminal]

See the edit to my last post.

So the difference between them ("CPL" and "Ferrari") is 2.7% on average...

 

[acoggan]

No, I wont.

But I will happily read what you link to and since you know where it is you can find it.

Working on it (seems to be longer ago than I recalled).

 

[acoggan]

So the difference between them ("CPL" and "Ferrari") is 2.7% on average...

Right...on average. The problem is that the accuracy of any such modeling is always limited by the accuracy of the input data, which is almost always ****-poor. That's why you need to look to individual cases, not summary stats.

 

[Ferminal]

Right...on average. The problem is that the accuracy of any such modeling is always limited by the accuracy of the input data, which is almost always ****-poor. That's why you need to look to individual cases, not summary stats.

I don't get it? The input data is exactly the same it has no bearing on the differences between two methods.

 

[acoggan]

No, I wont.

But I will happily read what you link to and since you know where it is you can find it.

Most of the action was back in 2009 and 2010...here's one particularly memorable quote:

"A sustained (over 40 minutes) power output of greater than 6.2 W/kg at the end of a Tour stage is simply not physiologically believable..."

And here are a couple of links to get you started...note that what's said in the comments are probably just as, if not more, important than in the blog posts themselves:

http://www.sportsscientists.com/2009/07/tour-2009-contador-climb.html

http://www.sportsscientists.com/2010/06/tour-on-horizon.html

 

[Netserk]

"A sustained (over 40 minutes) power output of greater than 6.2 W/kg at the end of a Tour stage is simply not physiologically believable..."

AFAIK only dopers have managed to do that, so I think it's spot on.

 

[acoggan]

I don't get it? The input data is exactly the same it has no bearing on the differences between two methods.

The average difference between the two estimates is similar to the average difference between either estimate and the criterion value...but in both cases, looking only at averages doesn't tell you how much/little difference there is in individual cases. A scatterplot as I originally post, OTOH, shows the latter rather well.

 

[acoggan]

AFAIK only dopers have managed to do that, so I think it's spot on.

That depends on how you define "doper". If you hold that someone is clean until officially sanctioned for doping, then no, it's not. OTOH, if you're willing to label someone a doper based on other criteria, then you've establishing a moving standard, and will always be right no matter what.

 

[Ferminal]

The average difference between the two estimates is similar to the average difference between either estimate and the criterion value...but in both cases, looking only at averages doesn't tell you how much/little difference there is in individual cases. A scatterplot as I originally post, OTOH, shows the latter rather well.

The absolute average difference between them against the criterion has nothing to do with the average difference between the two... You can scatterplot them if you really want to and one of the five will look evenly remotely close to the 45 line, and only two of the five are even in the right quadrant!

So, based on the data you have presented, there is a very noticeable difference between the two methods.

 

[acoggan]

The absolute average difference between them against the criterion has nothing to do with the average difference between the two...

Mathematically, no, but coincidentally, yes.

 

[Dr. Maserati]

Most of the action was back in 2009 and 2010...here's one particularly memorable quote:

"A sustained (over 40 minutes) power output of greater than 6.2 W/kg at the end of a Tour stage is simply not physiologically believable..."

And here are a couple of links to get you started...note that what's said in the comments are probably just as, if not more, important than in the blog posts themselves:

http://www.sportsscientists.com/2009/07/tour-2009-contador-climb.html

http://www.sportsscientists.com/2010/06/tour-on-horizon.html

This is what you wrote:

Originally Posted by IzzyStradlin
Power modeling is not detecting dopers, nobody is claiming it is

Actually, many take it as such, including Tucker himself, who just a couple of years ago proposed that anti-doping authorities use the data for that very purpose.

.........in the articles you quoted he does not specify that and is (as always) at pains to show that there are limitations.

You started a thread on a strawman and are proving something that no-one suggested.


From the 2 articles you posted:
http://www.sportsscientists.com/2009/07/tour-2009-contador-climb.html

I've hopefully managed to explain some of the other factors that must be considered in the Contador climb, but this question remains, without a doubt. It would be naive to dismiss it out of hand. I will say that performance analysis of single performances does not constitute proof of anything. In fact, it's a weak method of inferring doping. That was never my intention in yesterday's post, by the way (in case it came across that way). The better approach is to look at all climbs and work on averages, as I did for Tour winners from 1989 to 2001 in a previous post.

and:

And so to get the ball rolling, here is an interesting article that came out in New Scientist today, called "Superhuman performance could betray sport drug cheats". I know it because I'm quoted in it and helped with some of the analysis. The basic premise is this - in order to produce a certain power output on a climb, you have to have a certain capacity to use oxygen. The work done has an oxygen cost, and this cost tells you a good deal about the "ceiling". If the climbing power output predicts a ludicrously high oxygen consumption, then you have a waving red flag.

It's not proof, but a very suspicious question mark. And I can assure you, the Tour is littered with question marks, from the 1990s up to perhaps last year. This year's Giro d'Italia, many of you have already noticed, has seen a substantial drop in power outputs on the climbs, and they are now "physiological" again.

 

[Merckx index]

Because there are two aspects to the issue:

1) the accuracy of estimated power output; and

2) what we truly know and/or don't know about the limits of human performance.

Because the two are really separate, somebody could agree w/ me that estimated power is too inaccurate, but disagree w/ me about whether you positively identify doping based on directly-measured values.

In any case, the graph I shared really only speaks to the 1st issue listed above.

Which provides a good argument for gathering more SRM data, and releasing critical data that are already available (e.g., Froome pre-2011). If this were done, we wouldn’t have to have these discussions about which formula provides the best estimate of true power. We could move on to 2), which is a really important question.

But Froome/Sky, AFAIK, still haven't released data from his earlier years, or even explained why they won't. And it was either you or Alex (if the latter, I bet you agree with him), who went on and on about how it’s impractical to have SRM data for all the riders or even contenders in a GT.

If people won't release SRM data that are available, and make excuses about why more such data can't be obtained, is anyone surprised there are endless discussions revolving around estimates of power? Given how easy it is to beat doping tests, power outputs are the best indicators of how widespread doping may be that we have. The passport has shown that one can use fluctuations in certain physiological parameters from a baseline as highly significant indicators of doping. Performance ultimately derives from such physiological parameters, and has to have limits that are rooted in the limits of these parameters.

 

[Rip:30]

That depends on how you define "doper". If you hold that someone is clean until officially sanctioned for doping, then no, it's not. OTOH, if you're willing to label someone a doper based on other criteria, then you've establishing a moving standard, and will always be right no matter what.

Agree you can't really know the limits of "natural" human performance and the error in the power estimates is an issue, but I'd still bet that team Sky is juicing to the gills and that it will come out sometime in the future. Noting that Froome's performances are at least near that of Armstrong, Pantani et al. is not a statistically valid way to demonstrate he's doped up, but it's certainly interesting.

TL;DR in a decade +, Chris Froome will be also be a known doper, but the proof will not come from power output estimates.

 

[hiero2]

performance is never proof of doping, unless speeds are down then it proves cycling is clean. But when the speeds go up again it means everyone has moved on to advanced training plans instead of doping.

While coggan said it more succinctly - his response was insufficiently informative. You can't SCIENTIFICALLY draw the conclusion of "no-doping" because speeds are down. You can't LOGICALLY, OR scientifically draw the conclusion that when speeds go back up doping is not involved. The evidence and arguments don't support this as the only possible conclusion. Could be, but might not be.

OK Dr. Cog, I've taken a lot of good advice from your books and forum posts and whatnot over the past several years, but you and Hunter have gone totally off the rails over the past year or so.

Power modeling is not detecting dopers, nobody is claiming it is. It is simply providing a basis to compare climbing times, and therefore power outputs over the years. Ie, trying to compare racers who are riding the same climb in different races.

You know this, Hunter knows this. Claiming that people are trying to detect doping solely through these models is a strawman argument to disparage the work of some bright guys who don't happen to be certified coaches or physiologists.

One question remains, how do clean riders currently ride as fast as doped riders? Faster than historically clean(ish?) riders. Not a little faster, a lot. We have seen climbing times slow down due to improved drug testing, but now we've seen them surge again....how? Why?

Actually, many take it as such, including Tucker himself, who just a couple of years ago proposed that anti-doping authorities use the data for that very purpose.

Times and/or placements don't tell you enough already?

AC, actually, I think Izzy has this one right, and you seem to be missing the point, or disregarding the arguments. No, I don't think Tucker or anyone else has proposed sanctioning riders based on performance. Yes, it is being suggested as a guideline for testing, just like placement. No, placement is not enough - for a number of reasons. Placement over a period of years has more variables than performance. Also, a rider may perform in the "unreal" range - yet not place highly. Think Hincapie, or Porte, and you get the idea quickly. Also, you assume that high placement will mean a strict testing protocol, which, we have seen, may be assuming overly much. So, for many reasons, this could be a useful measurement tool.

 

[acoggan]

This is what you wrote:



.........in the articles you quoted he does not specify that and is (as always) at pains to show that there are limitations.

You started a thread on a strawman and are proving something that no-one suggested.


From the 2 articles you posted:
http://www.sportsscientists.com/2009/07/tour-2009-contador-climb.html

and:

Like I said, you need to read the comments section: Tucker explicitly endorsed the idea that power data be used by anti-doping authorities to detect dopers.

 

[acoggan]

AC, actually, I think Izzy has this one right, and you seem to be missing the point, or disregarding the arguments. No, I don't think Tucker or anyone else has proposed sanctioning riders based on performance. Yes, it is being suggested as a guideline for testing, just like placement. No, placement is not enough - for a number of reasons. Placement over a period of years has more variables than performance. Also, a rider may perform in the "unreal" range - yet not place highly. Think Hincapie, or Porte, and you get the idea quickly. Also, you assume that high placement will mean a strict testing protocol, which, we have seen, may be assuming overly much. So, for many reasons, this could be a useful measurement tool.

I must respectfully disagree. Estimating (or even measuring) power outputs in order to better target other anti-doping measures doesn't, IMO, provide enough incremental benefit to make it worth the effort and risk.

 

[acoggan]

Froome/Sky

Just to be clear: while you might have some specific individuals/teams in mind, my comments are purely generic in nature.

 

[Dr. Maserati]

Just to be clear: while you might have some specific individuals/teams in mind, my comments are purely generic in nature.

I read them - he does not explicitly (or even subtly) suggest that.

It says statements like this as an example.

...

And to respond, yes, absolutely. This exercise, and the one in the next post, is never intended to catch anyone, or prove anything, only to identify "flags", precisely because of the error, which we must always acknowledge!

I certainly hope that we see plenty of SRM data during the Tour, and we'll certainly do our best to give some thoughts on it!

Regards
Ross

 

[acoggan]

I read them - he does not explicitly (or even subtly) suggest that.

As I already pointed out up-thread, Tucker (and now others) have back-tracked significantly on their original positions.

EDIT: So after reviewing various sources (including my own 2-part rant on the general topic, posted to this forum), it appears that this back-tracking started when I first called Tucker out on his black-and-white stance (previously quoted verbatim) in a thread from 2010 titled "Superhuman performance could betray sports drug cheats".

 

[Almeisan]

When we seen Quintana's and Rodríguez hemoglobin values throughout the tour, even Mr. Coggan will admit it is proof of doping. But just the values without the performance that should suggest hemoglobin levels should drop down, it is no proof. Quintana and Rod could just have climbed in the gruppetto going easy without their hemoglobin dropping off.
But climbing faster in third week without hemoglobin dropping off, that together is proof.
Or imagine hemoglobin does drop off but performances do not?

Also, Coggan talking about ethics of the scientific process is a bit ironic. He and Coyle are basically scientific guns for hire to proof whatever a doped athlete wants to have proved.

 

[pmcg76]

As I already pointed out up-thread, Tucker (and now others) have back-tracked significantly on their original positions.

EDIT: So after reviewing various sources (including my own 2-part rant on the general topic, posted to this forum), it appears that this back-tracking started when I first called Tucker out on his black-and-white stance (previously quoted verbatim) in a thread from 2010 titled "Superhuman performance could betray sports drug cheats".

I would like you to explain why if Power outputs cannot be used as part of an anti-doping system, what do you see as successful weapons in the fight against doping. The entire history of the sport has shown that drug testing has proven ineffective against doping.

If you believe the current BP is more successful, what is the evidence that this is the case if Power Outputs/ VAM and climbing times are to be discounted as indicators of anything.

One thing that has puzzled me and still does is the climb to Aix-3-Domaine. Lots of people came out defending Froome when it turned out he had the 3rd fastest time ever on the climb. There were lots of excuses like more favourable conditions, first day in the mountains etc, etc but if we take the times as a whole for the stage, times are down on previous years despite all riders having the same favourable conditions as Froome which would back up the belief that the peloton is cleaner, but then we have Froome riding a time faster than most of the dopers. Any logical explanation for this anomaly.

 

[goggalor]

I see two massive outliers at around 400 W. Does each point represent one calculation accompanied by an actual power data file? Either way, on a Tour de France MTF you'll have a bunch of riders unevenly spread from ~450 W and down, so you should be able to spot these kinds of glaring calculation errors. I can't imagine anyone is actually thinking of sanctioning riders based on power estimates anyway.

 

[vrusimov]

I would like you to explain why if Power outputs cannot be used as part of an anti-doping system, what do you see as successful weapons in the fight against doping. The entire history of the sport has shown that drug testing has proven ineffective against doping.

If you believe the current BP is more successful, what is the evidence that this is the case if Power Outputs/ VAM and climbing times are to be discounted as indicators of anything.

One thing that has puzzled me and still does is the climb to Aix-3-Domaine. Lots of people came out defending Froome when it turned out he had the 3rd fastest time ever on the climb. There were lots of excuses like more favourable conditions, first day in the mountains etc, etc but if we take the times as a whole for the stage, times are down on previous years despite all riders having the same favourable conditions as Froome which would back up the belief that the peloton is cleaner, but then we have Froome riding a time faster than most of the dopers. Any logical explanation for this anomaly.

He was riding in his own exclusive atmospheric envelope!

 

[acoggan]

Coggan talking about ethics of the scientific process is a bit ironic. He and Coyle are basically scientific guns for hire to proof whatever a doped athlete wants to have proved.

**** off. I've never been paid to provide my opinion on any matter.

 

[acoggan]

I would like you to explain why if Power outputs cannot be used as part of an anti-doping system, what do you see as successful weapons in the fight against doping. The entire history of the sport has shown that drug testing has proven ineffective against doping.

If you believe the current BP is more successful, what is the evidence that this is the case if Power Outputs/ VAM and climbing times are to be discounted as indicators of anything.

My position has really has nothing to do with efficacy, but rather with principles.