You Can Run, but You Cannot Hide

At the 2015 Ironman Canada race, a grueling all-day triathlon consisting of a 2.4-mile swim, a 112-mile bike ride, and a full 26.2-mile marathon, the announced winner of the women’s 40-44 year-old age group category, Julie Miller, seemed to literally come out of nowhere to win.

Something about that result did not seem quite right to James Young, whose wife, Claire, was racing hard and ended up officially finishing third after a day-long duel with a racer named Susanne Davis.

Miller claimed to have lost her RFID timing chip, and was granted the victory by race officials, though neither Claire Young nor Davis had seen her on the course. When others later told Young there was no way Miller could have legitimately finished the race in the time she claimed, he turned to the Internet and began to pore over photos posted online by race photographers. Though her lack of a chip meant she had no official times at measuring points along the marathon course, called splits, she did appear in photos with other runners whom Young could identify by their bib numbers and official chip-measured times.

The evidence, Young said, was clear: given her position with the other runners he had found online, Miller would have had to run one of the two marathon laps at either a snail-like walking pace or at a near-world-record pace for a professional male runner.

Triathlon Canada officials subsequently negated Miller’s victory, as well as all her race results going back to 2013, for “breach of Triathlon Canada’s Code of Conduct,” and banned her from competition in sanctioned events for two years.

 

 Julie Miller competing in the 2015 Ironman Canada.
She lost her timing chip, shown at right, during the race.
Credit: FinisherPix.com

Young said he still will not accuse Julie Miller of cheating in that 2015 race, because the lack of a chip made 100% accuracy in analyzing her time impossible. However, the case illustrates the increasingly networked world around high-profile races, and a growing community of arbiters ferreting out cheaters through race websites and social media. One man has even developed an analytic tool that reveals people who finished the iconic Boston Marathon in times significantly slower than those they claimed to have run in races to qualify for Boston, a starting point into investigating whether they cheated their way into the race.

There is more than pride on the line. Marquee races bring in millions of dollars from sponsors and competitors who travel to get there. Increasingly, they are also gold mines for charities; the Boston Marathon, for instance, reported participants running for charity raised $28.3 million in the 2015 race alone.

“It’s a massive draw,” Young said. “Just from a business perspective, if those races don’t police and don’t maintain credibility for people getting those spots, it’s going to kill what they are able to offer. If Boston doesn’t get a grip on how to stop these bogus people getting spots, if Ironman can’t get a grip on ensuring the age groups are fair, there will be more people cheating, and when more people are cheating, then more of the others will say, ‘what’s the point of even trying?’ I think the onus has to be on the businesses themselves to say, ‘we’re going to get stuck into making sure people are honest.'”

As Young’s detective work demonstrated, the ubiquity of Internet resources surrounding races can make ferreting cheaters out possible, if time-consuming. “It takes a certain kind of fairly analytical person like me, using computers, who’s using that sort of logical thought process, and also I had enough time—it took hours and hours—and I had a vested interest in it.”

For in-race cheating situations, Young said it might be plausible for race timing app developers to create tools that could help race directors find inconsistent results.

“In theory, when you have these apps, you have the start time, the time they came out of the swim, then the time they’re in transition, the time they got on their bike, and various splits on the bike,” he said. “It’s effectively just a database of the triathlete throughout the day. You could correlate that database and almost plot them on a graph going along. And you could plot several people and you could look for wild variances.”

Cheating Before The Race Starts

While one of the most publicized pre-Internet cases of race cheating happened during the 1980 Boston Marathon—a runner named Rosie Ruiz jumped into the race in the last mile or so to break the finish-line tape first for female competitors, and was quickly found out—much of the cheating around the race today happens before people even get to Boston, and in finishing spots far from those who would get a winner’s trophy. Some people, who likely could never legitimately meet Boston’s strict qualifying time, try various methods of creating false times in qualifying races, including failure to complete race courses in qualifiers, or somehow enticing faster runners to wear bibs officially registered in the scheming runner’s name.

Enter Derek Murphy and his Marathon Investigation project, which demonstrated a computational way to help catch race cheaters after the 2015 Boston Marathon.

Murphy, a Cincinnati, OH-based business analyst, developed an algorithm that reveals which runners finished the Boston Marathon 20 minutes or more slower than their time in their qualifying race (those times are available on sites like Athlinks.com). Boston issues its bib numbers based on qualifying times, which made it easy for Murphy’s algorithm to accurately predict the true pace a given runner ran at his or her qualifying race (an accompanying scatter plot includes 2015 Boston runners marked in orange whom Murphy flagged as having invalid results with what he said was 100% accuracy).

In total, Murphy reviewed 1,409 runners as outliers, and deemed 12 results invalid as a result of cutting the qualifying course, 33 invalid through a bib swap, and 11 that either had someone run the qualifying course for them or somehow got into Boston because race results were modified somehow. Those runners were reported to the Boston Athletic Association, the race’s organizer (which did not respond to requests for comment on this story)

Murphy’s approach is suited to the large number of runners in the middle of the pack rather than first-place contenders, but in the running community, cheating is reviled no matter one’s finishing time, especially when a shot at running a race as meaningful as Boston is at stake. His algorithm, he said, helps winnow out those who have taken a coveted spot for which someone else may have legitimately qualified.

“It’s stuff that people may not notice,” he said. “Who’s going to notice somebody who barely gets through a qualifier in time?”

Plenty of people have noticed, it turns out. Finding imposters could have significant ramifications for those who did the right thing and still could not run; Boston race officials had to turn away more than 4,500 runners who qualified for the 2016 race due to a capacity field of 30,000.

Murphy’s platform is rapidly gaining credence among the running community. Prior to early April, when Murphy ran a post that explained why a popular running blogger was found to have swapped bibs to gain entry into the 2016 race, his site had about 40,000 visitors. By the beginning of May, two weeks after the Boston Marathon, he had over 325,000, receiving almost 20,000 hits in one day (May 4-5), and getting a steady stream of tips about dodgy results and photos of those who ran Boston using counterfeit bibs.

Murphy said he has not received any firm offers from race directors to partner in ensuring their entrants are all honest, “but I have found they have been very responsive.

“I have not pushed that at all. I have been making some great connections. My plan for the short term is to keep getting my work our there as it relates to 2016, and then start pushing towards making this viable as a potential full-time venture in some form as I try to get myself firmly established as an authority on the topic.” 

Gregory Goth is an Oakville, CT-based writer who specializes in science and technology.