Become a Member

Get access to more than 30 brands, premium video, exclusive content, events, mapping, and more.

Already have an account? Sign In

Become a Member

Get access to more than 30 brands, premium video, exclusive content, events, mapping, and more.

Already have an account? Sign In



The Outer Line: Researchers seek the power to believe pro cyclists

Researchers analyze power data to build a new weapon in the fight against cheating in pro cycling.

Don't miss a moment from Paris-Roubaix and Unbound Gravel, to the Giro d’Italia, Tour de France, Vuelta a España, and everything in between when you join Outside+.

Chris Froome’s raging comeback in the Giro d’Italia has been the talk of cycling fans everywhere and has ignited fierce debate in the scientific community. Fans of the sport were electrified by his exploits; but as some sports scientists estimated the power and wattage output related to his stunning attacks on the Zoncolan and Bardonecchia stages of the race, many of Twitter’s established cycling critics raised questions about Froome’s physiology or outright questioned the legitimacy of his athletic gifts.

Twitter is, of course, on fire with all manner of accusations and insinuations as to how Froome could possibly beat the math and win the race. Nor did it help Froome or Team Sky that he did not make his power meter data immediately available for review. (On Monday, Velon published some of Froome’s power data from the Giro’s key moments.)

Even his fellow pros were a bit dubious. George Bennett (LottoNL-Jumbo), in the heat of the moment, drew an off-the-cuff comparison to Floyd Landis’s near-impossible-but-doping-enabled comeback at Morzine in the 2006 Tour de France. But, even though the speculation and conspiracy theories will go on endlessly, Chris Froome wears the maglia rosa, and at least so far there is no evidence to suggest that he won it unfairly.

But what if — instead of all these insinuations and theoretical calculations — there was a way to truly demonstrate and document “outstanding performance” versus “performance enhancement?” What if there was a surefire way to prove, once and for all, when we can truly believe in what we are seeing, and when we might want to be more skeptical? It is encouraging in this regard to report that a team of scientists has recently made progress on just such a solution. Endurance sports like elite cycling may soon have a new weapon to reinforce sporting integrity.

Dr. Michael Puchowicz of Arizona State University is not exactly a household name in cycling, although that might soon change. A former competitor himself, Puchowicz is a researcher in the exercise physiology field and has consulted to private team anti-doping programs and an athlete certification program called Clean Protocol. Puchowicz was the lead author on a review article entitled “The Critical Power Model As A Potential Tool For Anti-Doping,” which comprehensively discussed issues relating to using the critical power model to identify cheating in endurance athletics. While most cyclists are familiar with “power” as a training tool, and power meters made by various manufacturers have become ubiquitous in the sport, Puchowicz and his colleagues suggest that power meters may also soon become a valuable tool for catching potential dopers.

These researchers describe critical power, or CP, as a threshold for purely aerobic exertion which an athlete can sustain for an extended length of time. A second measurement for “work-above-threshold,” called W’ (“Work Prime”), measures the amount of effort an athlete can exert in excess of their CP line. As described in scientific detail in the paper, researchers can record athlete performances in high-intensity exercise, map the specific curve of an athlete’s CP and W’ levels, and measure the ability for that athlete to exceed those personally-defined limits for some period of time. Previous studies have proven that these measurements for CP and W’ are stable in highly-trained elite athletes and diminish with age. Critically, these studies also indicate that abnormally large improvements in measured CP and W’ levels may only be achieved by employing some sort of performance-enhancing drugs.

A measurement model for CP has been around since at least the 1960s and was refined in the early 1980s. CP models have been proven in whole body exercise studies, animal studies, and routinely used with cycling time trial to exhaustion tests. “We can accurately track this in any athlete,” says Dr. Puchowicz of the model. “You can measure these power points accurately in a lab test, and measure them in real-world scenarios of the sport.”

The goal of doping is to enhance performance — doping is the cause; enhanced performance is the effect. This is why the authors focused on analyzing performance rather than just biological analysis in their approach to anti-doping. The actual performance data provides important and independent markers needed to help catch cheaters. As examples, the authors point out that when EPO use started in pro cycling in the late 1980s, there was a period of rapid and unlikely improvement in individual and peloton speeds in the world’s most important races. But those performances began to stagnate and decline after 2004 as better tests were implemented, first with the EPO urinalysis in 2002, and again after 2008, when the UCI started the biological passport to catch riders who tried to re-infuse their own red blood cells.

As the blood data showed for pro cycling, performance changes marched in lock-step with the prevailing doping methods in favor at the time. But ironically, the advent of power meters in cycling during the same time frame has enabled parallel, accurate tracking of doped performances to the point that performance itself has become a more and more reliable indicator of the potential cheating. Cycling’s coaches and serious riders today are very familiar with power meter analysis measurements like mean maximal power (MMP) profile or “record power profile.” As the paper’s authors state, these different profile views of a rider’s performance data are “predictive of future performances and evolve as the athlete develops over time, such that unrealistic increases in the powers sustainable (over) the indicated durations could serve as evidence for doping.”

The review that Dr. Puchowicz and his colleagues have recently completed, to be published shortly as part of a series on human performance in the journal Frontiers in Physiology, defines the steps necessary to implement a robust and standardized critical power model which can indicate if an athlete may have doped. “We’ve already gone down the path of looking at other (indirect) markers,” says Dr. Puchowicz about current anti-doping methods. “We’ve arrived at a point where we have to propose another measure that’s independent of biological tests, especially now that we know how sophisticated some cheating has become.”

The “sophistication” which Dr. Puchowicz and many other researchers are pointing to is the ongoing arms race between cheaters and anti-doping authorities. A growing consensus, especially after revelations in the documentary “Icarus” and through the work of investigative journalists, is that doctors who are “preparing” athletes with doping methods are always a few steps ahead of the anti-doping scientists trying to catch them today. As related by the paper’s corresponding author, Dr. David Clarke, a sophisticated doping program can sidestep the biological passport by micro-dosing over time to keep the athlete’s hormone and blood markers elevated, all while avoiding traditional detection by current passport data trending and analysis models. In this light, a properly implemented performance-based model becomes an important independent indicator of doping (and the type of doping) when medical sophistication can fool the authorities.

From their review of many previously published scientific athlete studies, the team built comprehensive tables of averaged measurements taken from hundreds of elite and highly trained athletes. But the research team went further and built corresponding tables of athlete data from previous studies which measured the actual boosting effects from steroid and blood doping enhancement methods. The team then mapped simulated doping improvements against their known athlete averages. The resulting proposed model sets a benchmark for how anti-doping scientists can be tipped off to the type of doping an athlete may have employed, when the doping window may have started, and from this more accurately review the athlete’s complete case history.

Dr. Puchowicz is quick to point out that while measuring power output will make the biological passport more effective, the CP test should not be used on its own to prove a case. A flagged CP finding and a flagged biological passport finding, when taken together, are stronger proof of cheating than either finding viewed alone. So, when an athlete’s CP is measured in a range which predicts an enhanced or doped performance, targeted testing or the athlete’s biological passport from the same time period can be more meticulously reviewed for possible doping manipulation. “We can look at an uptick in the athlete’s CP measurements as captured by a power meter, compare that to the biological passport, and vice versa,” says Dr. Puchowicz.

Currently, the athlete biological passport is focused on things which are likely only detectable in the time leading up to races when doping “preparation” is occurring; on the other hand, performance enhancement is readily apparent when the racing is actually happening. And whereas there are known (and yet to be revealed) methods for subverting biological doping tests, CP is proven to be largely insensitive to manipulation when the data collection is accurate. Given that the techniques for measuring CP are well established, say the authors, “The underlying framework for an anti-doping (athlete) performance test already exists, such that future developments in analytical approaches should be reasonably straightforward.”

There is already a published example of critical power independently proving doping. Sergei Iljukov of the Research Institute for Olympic Sports, Jyvaskyla, Finland, and his colleagues documented a case study in competitive running and proposed a new “Athlete’s Performance Passport.” In that study, published in March, the scientists modeled the CP data of a group of distance runners and marathoners based on their past performances. One of the runners showed clear and unusual deviations from his predicted performances in future events. Targeted tests and the athlete’s biological passport markers, when compared to the independently measured CP observations, led to an adverse anti-doping finding — and a suspension.

However, it may be several years before critical power is an accepted anti-doping measurement in cycling. There needs to be a standard measuring tool defined for power, and as many cycling consumers know, the market is filled with different power meter brands. Some measure it at the hub, others from a single crankarm, bi-laterally with pedals, and others at the crank spider. And Dr. Puchowicz points out the need for scientists to take the next step for cycling and perform “real-world analysis, with experimentally vectored doping and placebo groups from which we can verify our predicted results.”

Other variable factors include the old argument of effort validity — that is, is what we observe in training the same as how the athlete will perform in the heat of a highly competitive race? Cycling, except for velodrome racing, is a very dynamic sport with every course being a little bit different, and riders are rarely in peak condition year round. For example, a track pursuiter’s best splits are usually faster in a World Cup than in mid-summer when they are not in track-specific form. Even for known and famous climbs, like Alpe d’Huez, an athlete racing up it just once in a time trial would have different results than a rider tackling it from a breakaway after four hours of racing. The authors of the paper acknowledge these weaknesses and other potential errors in their analysis of the future data collection methods.

And then there is the question of who would administer a kind of “athlete performance passport.” For it to be a truly independent reference and comparison marker, should it be a WADA program or an arm of a research institution? Or could it be administered by a consortium of national endurance athletics organizations, representing sports like swimming, cross-country skiing, athletics, and cycling? The only certainty seems to be that with more examples of superhuman efforts defying conventional logic in endurance sports, new tools for reinforcing athletic integrity can’t be introduced soon enough to answer the big questions and quench the doubts of skeptical observers.

However, Dr. Puchowicz and his colleagues like Dr. Clarke are confident they are on the right path with their framework to implement the CP model as a strong but independent reference to reinforce the current biological passport. There is no doubt that critical power is highly sensitive to manipulation by PEDs, and they believe that this first step of presenting a fully dressed data model example will eventually lead to more research, testing, and a validated protocol for power measurement in anti-doping. The distance running case example has been the first application of one such model in the real world, and Dr. Puchowicz believes there will soon be others. “I think we can do a wider validation study in about a year,” says Dr. Puchowicz of the team’s future research. “Running was the low hanging fruit, but sports like cycling are next. We have more work to do, but with our modeling framework now in place, we can get there.”

Editor’s Note:The Critical Power Model As A Potential Tool For Anti-Doping” was a collaboration of four independent researchers: Dr. Puchowicz, Dr. Nathan Townsend, Dr. Michael Koehle, and Dr. Clarke, with assistance from co-authors Assaf Yogev and Eliran Mizelman. Readers can sign up for email notification when the full paper is published shortly.