Swings in betting odds suggest an average of 23 professional tennis matches might be fixed each year, according to a study that found suspicious patterns in three matches at Wimbledon.
The research analyzed 6,204 first-round matches on the men's and women's tour between 2011 and 2013. The study, written by Ryan Rodenberg, an assistant professor of Sports Law at Florida State University, and South Bend, Indiana-based professional tennis gambler Elihu Feustel, was published by the Journal of Prediction Markets.
The study tracked betting market prices against the "correct" price determined by two predictive models. The methods use data from previous matches to judge the likelihood of each player winning. In 20 cases, the market price drifted away from one model's price by between 16 percent and 29 percent before play started, the study found.
Such a deviation in pre-match odds could have resulted from about $100,000 of wagering on a lower-ranked player winning at eight sports books, the study said. While the study's methods have limitations, they are "a strong starting point for detecting and predicting match fixing," according to the authors.
A fixing scandal at Wimbledon could scare off tournament sponsors and hurt sports gambling companies, who generated a total of $58 billion of gross profit in 2012, Warwick Bartlett, chief executive officer of Isle of Man, U.K.-based Global Betting & Gaming Consultants, said by phone.
"It would be a problem because credibility is everything in sports," Bartlett said. "People don't want to bet on roulette if the wheel is rigged."
Some of the most unusual swings happened in three first- round contests at Wimbledon in 2011 and 2012, and one at the 2012 London Olympics. The study didn't identify the dates of the games or players involved. It also found opening-round matches at the 2011 French Open and 2012 Australian Open with unusual betting patterns.
Wimbledon spokesman Johnny Perkins referred a request for comment on the study's findings to the Tennis Integrity Unit, or TIU, which declined to comment. The unit's spokesman Mark Harrison cited a policy of not discussing cases because they might lead to rumors and innuendo about innocent players. Australian Open spokesman Darren Pearce declined to comment, citing the same policy as the TIU.
Emilie Montane, legal director of the French tennis federation which runs the French Open, said the tournament has had an algorithm made by Sportradar AG since 2010 that flags unusual betting patterns.
"We have had small alerts very few times, and they were all explainable by sports reasons" such as injuries, Montane said by phone from Paris.
To be sure, matches with irregular betting patterns aren't necessarily manipulated. A possible explanation for an unusual swing could be tour insiders placing bets or leaking information to a gambler about a player's injury or illness without the athlete's knowledge or consent.
While that's against tennis rules, it's hard to police, according to Rodenberg.
"It's incredibly difficult for tennis authorities to make rules that are binding for physios and coaches," Rodenberg said.
Rodenberg said he conducted the study to find out how much progress tennis had made since a 2008 report commissioned by the sport's authorities found that 73 matches in five previous years had "suspected" betting patterns, and 45 warranted further review.
"Our results are fairly similar, they are in the same ballpark," Rodenberg said.
Fixing is more likely instigated by players "struggling to make ends meet" than criminal syndicates because the low volume of first-round betting in tennis doesn't make it worthwhile for organized crime, according to Bartlett.
After the 2008 report, tennis authorities set up the integrity unit to investigate possible cases of match manipulation. The London-based unit's website lists nine players who have been banned for fixing or other betting-related offenses since 2011.
In the latest case in June, authorities banned Russian player Andrey Kumantsov for life after finding him guilty of 12 charges between 2010 and 2013. Kumanstov, who didn't play in the main draw of a Grand Slam during those years, had a career-high rating of No. 261. Efforts to reach Kumanstov through the Russian tennis federation, which said it didn't have contact information for him, were unsuccessful.
Feustel, the study's co-author, said in an interview earlier this year that he has data from 260,000 tennis matches dating back several years which he has used to create an algorithm to make bets.
The ELO model used in the study gives each player a rating that is used to calculate the probability of winning. The other model estimates that probability based on service and return points won by each player against common opponents.