Here's a recent article by Tom Perrotta and Carl Bialik talking about how the official numbers for UE's at Wimbledon look too low.
It's Very Hard to Make a Mistake at Wimbledon
Welcome to Wimbledon, where the lawns are lush (at least near the net) and the tennis is a little too perfect.
You see, at Wimbledon, players don't appear to make many mistakes. In the last four years, including the first four rounds this year, about 24% of the points played in singles at Wimbledon ended with an unforced error. According to the tournament, this occurs when a player is "not judged to be under physical pressure as a result of the placement, power or spin of the opponent's stroke." Ultimately, though, "It's a judgment call," said Keith Sohl, a consultant for IBM, which trains match scorers and oversees the tournament's stats.
The scorers at Wimbledon have a reputation for being generous. Take Novak Djokovic's third-round contest against Jeremy Chardy. Djokovic was assigned just three unforced errors, and none until he double-faulted late in the third set. Tim Henman, commenting on the match for the BBC, was skeptical of Djokovic's perfection. "I'm not having it," he said early in the third set.
Nor should he. In the fourth game of the match, Djokovic hit a backhand into the net and threw back his head, clearly annoyed that he had missed. It wasn't scored an unforced error. Djokovic didn't even have to move to hit the shot, though the ball did bounce high, which is something scorers could have deemed a quirky grass-court bounce. "The surface can force errors that wouldn't be errors on a hard court," Sohl said. In all, we found 11 unforced errors, eight more than the scorers.
In Agnieszka Radwanska's 7-6(5), 4-6, 6-2 victory over Li Na on Tuesday, the scorers gave Radwanska 18 unforced errors; we counted 23. Then there was Roger Federer's last shot in his second-round defeat, a routine backhand miss, which wasn't deemed an unforced error. For the seven-time champion, perhaps it was a parting gift.
Overall, Federer was marked down for just one unforced error in the fourth and final set of his loss to Sergiy Stakhovsky, though the Journal counted six.
Where the Journal and the officials deviated, it was almost entirely in one direction. Eight times official scorekeepers spotted an unforced error when the Journal didn't, compared to 67 shots that went the other way.
There were no common threads to explain why some shots looked like unforced errors to the Journal but not to the official scorekeepers. About half were forehands, half backhands. Just a few were volleys and a handful of others were service returns. They came early and late in sets. If there was a common thread, it was that players were treated kindly if they were going for a tough shot—even if they didn't have to.
It's difficult to directly compare the majors for how they score errors, because not all of them score winners and errors for every professional singles match. To approximate the differences, we looked at the men's and women's singles semifinals and finals for the last four competed major tournaments, since all of those matches do have scorekeepers. We calculated, for each major, the number of points that ended in errors which required judgment calls: those not won by winners or lost by double faults. Then we calculated the percentage of those that were scored as unforced errors.
The results were definitive: Wimbledon's scorekeepers are the most generous. Just 30% of non-double-fault errors at the Wimbledon semis and finals last year counted as unforced, compared to 47% for the equivalent matches at last year's U.S. Open, 51% at this year's French Open and 55% at this year's Australian Open.
There are legitimate reasons why more errors are forced on the grass at Wimbledon. The courts are the quickest of any Grand Slam, robbing players of time. More points are played at the net, which forces opponents to go for low-percentage shots. And four of the six semifinal and final matches at Wimbledon last year included Serena Williams and Roger Federer, two players whose aggressive styles force the action, and lots of errors. There is also evidence that other majors, at times, give players the benefit of the doubt, such as with an oft-repeated stat about Fernando Gonzalez's semifinal defeat of Tommy Haas at the 2007 Australian Open: that Gonzalez hit 42 winners and three unforced errors. Those three unforced errors were "a little bit of a myth," according to a recent revisit of that match.
But those factors probably can't explain all of the discrepancies. The Journal's scorekeeping suggests another factor is relatively generous scoring at Wimbledon.