Zwift suspended two riders from eRacing for data discrepancies that suggested "data fabrication or modification."

The virtual riding platform, which surged in popularity in 2020, is trying to institute protocols to prevent data manipulation for profit.

Zwift's latest communication on digital fraud concerns Lizi Duncombe and Shanni Berger. In both cases, Zwift made a decision to sanction the riders from competing on the platform for a period of six months retroactive to the date of the alleged violations.

With digital racing becoming a venue for meaningful competition with prize money, the possibility of file manipulation was perhaps inevitable. The first UCI eSport World Championship is scheduled for December, and Zwift has stated that it will adopt a zero-tolerance policy.

Certain eSport race organizers have stipulated that power data from a rider's smart turbo trainer must be verified with a secondary power source such as a power meter. By comparing the two files, platforms (and organizers) can identify poorly calibrated trainers and power meters.

Duncombe's investigation was triggered by her fourth-place finish at the Zwift Racing League Women's Qualifier event on September 17; Zwift had requested reference files from riders, and a close examination of that data revealed a series of discrepancies were revealed.

According to the board decision released by Zwift, Duncombe originally submitted 90 seconds of warm-up recordings as validation data and then did not fully dual-record the race, even though she determined this to be a mistake and that the dual recordings were not viable. decision to revoke the results of the race based on the fact that it was not.

Duncombe then provided another dual recording of a Favero Asioma power meter pedal paired with a Garmin Edge 820; Zwift contacted Duncombe and she responded:

"The first [dual recording file] , was a warm-up file saved before restarting the Garmin. When I charged the Garmin after the race, the race file I sent was uploaded a few hours later and I was able to record the entire race before the battery ran out."

She also reported:

"The first time I ran the race, I was able to record the entire race before the battery ran out.

Initial inspection of the data appeared legitimate, but further inspection highlighted anomalies.

The double-reference file that Duncombe sent to ZADA (Zwift Accuracy and Data Analysis) contained a .FIT file tagged with a version ID of "562." 562 is the version found in the FIT file created by Zwift ID, and the Garmin Edge 820 computer uses "1250" as its version ID value. This suggested to ZADA that the file had been manipulated.

In addition, her support data file had significant timing discrepancies as well as GPS data that could not be read by the Garmin device. Finally, the data correlated perfectly with the in-game Zwift records (correlation coefficient of R=1.00).

Duncombe's subsequent response included the claim that the Garmin was paired with a turbo trainer rather than pedals and the suggestion that the data was corrupted. Both claims were denied by ZADA.

Berger's case arose after she finished second in the Off the MAAP women's race; according to the Zwift release, the initial revocation was based on the fact that, despite the event's stipulation that riders use turbo trainers as the primary power source for the race, Berger due to his use of a power meter. Furthermore, there is a 0% variance between the two records, suggesting that the second file was in fact a duplicate of the original

.

This suggests that Burger had connected his power meter to both the Zwift device and the Garmin computer. Although a simple mistake, Burger's result was cancelled because he could not double record the results.

In response to the revocation, Berger's team manager and team owner provided the Zwift logs. The txt file claims that the power supply was in fact a Saris Smart Trainer.

However, the same log.txt file was also stored on Zwift's server, and upon cross-referencing it, ZADA discovered that the power supply was a Stages power meter, along with 26 other inconsistencies, suggesting that the log.txt file from Burger had been edited.

Burger repeatedly denied any wrongdoing, and both she and her family contacted Zwift, stating that she was innocent and did not have the technical ability to manipulate the file. However, according to Zwift, Berger later sent an email "apologizing for any misunderstanding" and stated that "it is very possible that I [Berger] made a mistake in the software due to human error."

Both athletes can still use the Zwift platform, but it has been suspended until March 19 and February 17, 2021, respectively.

Cycling-related highlights from today's Black Friday sale. Check out our guide to the best Black Friday bike sales, where you'll find everything cycling-related, including bikes, helmets, clothing, and more.

Today's Deal: USA

Today's Deal: UK