Can performance be used as an indicator of doping?

While cycling is supposedly going through a “new era” of clean performance, fans are still jaded by the lies and deceit of the recent past. Team Sky has come under heavy scrutiny, not because of anything in particular the riders have done wrong, only because they’ve shown such dominance.

Heading out the door? Read this article on the new Outside+ app available now on iOS devices for members! Download the app.

While cycling is supposedly going through a “new era” of clean performance, fans are still jaded by the lies and deceit of the recent past. Team Sky has come under heavy scrutiny, not because of anything in particular the riders have done wrong, only because they’ve shown such dominance.

After Sky’s remarkable performance on stage 8 of this year’s Tour de France the critics have come out and scrutinised Froome’s win. Without access to direct power data, the critics have had to rely on mathematical modelling, climb times, elevation data and a handful of assumptions in their attempt to quantify Froome’s performance and speculate about whether he has doped or not.

But can performance actually be used as an indicator of doping? Of course climbing up Alp d’Huez in 30 minutes raises red flags, but these aren’t the types of margins we’re dealing with.

VeloNews’s Andrew Hood did an interview with David Brailsford last week in which he asked the Team Sky principal why the team’s riders won’t release their power data to the public. Brailsford replied:

“There is so much pseudo-science out there right now. If you release the data, there are very few people who can properly interpret and understand that data. All you’re going to do is create a lot of noise for people who are pseudo-scientists. You can even write magazines about it.”

Brailsford is referring to “Not Normal”, a magazine released by French journalist and former Festina trainer Antoine Vayer. Released just before the Tour de France, “Not Normal” took 21 of the most successful riders from LeMond to Armstrong to Evans, quantified their performances and then ranked them across an index of suspicion.

The three categories Vayer used were “suspicious” — a power output of 410 watts at threshold — “miraculous” — above 430 watts — and “mutant” — above 450 watts. Note that Vayer standardises the performances of riders against an “average” rider weight of 70kg, allowing him to compare the performances of heavier and lighter riders on the same scale.

Vayer wrote a couple days ago that he calculated Froome’s power output on the stage 8 Ax-3-Domaines climb as 446 watts (scaled to a 70kg rider; the equivalent of 6.4 w/kg for Froome). In the article for Le Monde he described Froome’s performance as “miraculous”; on the phone to me he labelled it “not human”.

The question here is, can performance be used as an indication of doping? Can we say that if a rider is pushing above a certain number of watts they’ve exceeded human physiological limits and that they’re doping? I asked Vayer that very question.

“I don’t use the words ‘doping’ and ‘clean’. I don’t say if a rider is either. I just show people the numbers and let them make their own decisions.”

And as for Vayer’s motivations?

“I want to kill doping. I think my work brings attention to this problem. People who are not even cycling fans recognise me on the street and congratulate me on my work.”

I asked Klaas Faber, Dutch anti-doping expert and statistician, about his thoughts on performance being used used as an indicator for doping:

“Compare it with hematocrit. Riders could boost it up to 50% with impunity. It’s the same with power. With the difference that power can be monitored on the bike. Power is just another radar and a lousy one.”

That is, you can do whatever you like as long as you’re within the limit. And just because you’re under the limit doesn’t mean you’re not doping.

Ross Tucker is and exercise physiologist in South Africa who co-writes, a well-written, thought-provoking blog. Tucker focuses on performance analysis by calculating VAMs (vertical metres of climbing per hour) and watts per kilo (w/kg) of riders during the Tour de France.

After Froome’s win on stage 8 he calculated that the rider’s power up Ax-3-Domaines was 6.3w/kg or 6.5w/kg, depending on which model you use. (By way of contrast, the magic number Dr. Ferarri used to guage Lance Armstrong’s Tour winning form was 6.7w/kg, at threshold.)

Tucker is very careful to acknowledge the shortcomings of the modelling he uses and puts the numbers into context by stating variables such as how far into the race the effort took place, the tactics at play, wind direction and so on.

He also cautions people to step back and watch what happens during the Tour as a whole instead of focusing on one performance. As we saw, Sky paid a physiological price for their efforts on stage 8, with many fading the following day.

Despite the constructive skeptiscism and insights that Tucker provides, I see many people skipping past his caveats and quickly assuming that if an athlete pushes more than a certain number of watts per kilo then he is definitely doping.

Tim Kerrison is a former Australian and British swim coach and now Head of Performance at Team Sky. I spoke with Tim to get his thoughts on the issue. While you might think Kerrison has a conflict of interest in this area, given his involvement with Sky, there’s no doubt he’s an expert in the field of human physiology and performance:

“There are a lot of assumptions that go from observing a performance and translating that to a power-to-weight quantity. And then there’s a lot of assumptions made of what physiological capability is possible or impossible with a cycling athlete without doping.

It’s a worthwhile discussion to have that raises a lot of issues about performance, physiology, and limitation, but I think it’s a very long stretch to go from how fast someone rides their bike up a hill, to whether or not they’re doping.”

As mentioned, there are many factors that make it a stretch for someone at home to calculate a rider’s power-to-weight ratio, just by timing how fast they climb and how heavy the rider is.

Tim Kerrison explains that the variables include:

” … the gradient of the climb, if they’re climbing solo or in a group, the temperature and humidity, the wind direction, where that climb occurs in a race. And there’s a big difference between a climb that goes from sea level to 1000m, versus one that starts at 1000m and goes to 2000m in the power the athlete can produce. It could be the exact same climb, but depending on the altitude the athlete will produce two very different power outputs.”

While Dave Brailsford isn’t in favour or releasing his riders’ power data to the general public because of the noise it’ll create, Kerrison offers a slightly different view:

“At the end of the day, I’d love for information to be in the hands of experts, who will apply scientific rigor to the analysis to understand the data and the many factors that affect it.”

If this data is going to be analysed with due scientific rigour and potentially raise flags, the instruments being used (i.e. powermeters) need to be reliable, accurate and consistent. There are a range of power meters on the market and all of them have their own way of calculating power. All of them have their own issues with reliability and consistency as well, and they’re not always calibrated properly.

Kerrison says, “It’s a continuous challenge for us to keep all our equipment calibrated and accurate all the time. And like any scientific measurement, it’s a very important aspect to make sure that the information and data that other people are interpreting is valid. “

I asked the late Dr. Aldo Sassi about this same issue back in 2010, and his sentiment was the same:

“Too high of a performance can be an element to suspect unphysiological performance and doping. The problem is how to measure the performance through biomechanics. You have to make estimations, and do to an estimation you always have 3-4% error. If instead you use a power meter, you need to check the calibration of each rider. Otherwise you could have the same errors in the measurement.”

So if there are concerns about the reliability of the performance data being captured, it’s only fair that we should be extremely careful about judging performances based on that performance data.

It’s been suggested that ~6.2 w/kg (at threshold) is the limit of human physiology when it comes to cycling. But Tim Kerrison isn’t convinced about the idea of a hard limit on performance.

“I don’t buy the concept that there is a relationship between measuring power and physiology. Measuring power is not measuring physiology. It’s output. It’s a big step to go from measuring power to the physiology of an athlete.”

Instead, he takes a more ambitious view.

“How many times in history have people said that things aren’t possible? Then 10, 15, 100 years later it’s commonplace. So many things people say are physiologically not possible, but that’s not a winning mentality to me to put limitations to human performance.

What we aspire to do as a team with our athletes in the long-term is to surpass those limits that people are saying are physiologically impossible — and I’m sure people on other teams are aspiring to do those things as well. It’s a pretty shortsighted approach to think there’s a limit to human performance.”

To Kerrison, it comes down to the efficiency of the human machine.

“The human body is a machine that’s frankly not very efficient. It’s somewhere around 20% efficient. So it only takes a small evolution of training practices, or a particularly efficient human specimen to go from 20% to 21% (for argument sake). It’s only 1%, but results in a 4 or 5% increase in power output. So there is a large margin of error there to go from the power an athlete can produce, to the physiology of an athlete.”

There’s also the fact, says Kerrison, that exceptional performances are just that.

“The best performances in every sport are outliers. So the fact that you’re an outlier alone, and the fact that the performance is exceptional alone, is not evidence that somebody is doping. It has to be a lot more than an absolute value of the data, and a lot of thought needs to go into how that data will be used in that context.

So can we draw a line in the sand and say one performance is legitimate and one is doped, based purely on performance data? In my opinion, from the discussions I’ve had, using direct or derived performance figures is an interesting exercise and debate, but is by no means a definitive indicator of doped performances. There is a lot of context missing when looking at a single number and a lot of assumptions that are behind these numbers.

Cycling has certainly taught us that a healthy amount scepticism is good, but for now, wait for more substantial proof before raising the red flags. Sit back and enjoy the rest of the Tour de France and don’t get too bogged down in the numbers game.

[ct_highlight_box_start]Thank you to Tim Kerrison, Klaas Faber, and Antoine Vayer for their views on this topic. Dr. Michael Ashendon did not wish to comment and we contacted Dr. Ross Tucker but did not receive a reply. [ct_highlight_box_end]

Trending on Velo

An American in France

What’s it like to be an American cyclist living in France? Watch to get professional road cyclist Joe Dombrowski’s view.