Beware the abuse of statistics in sports

Feb. 11, 2013, 11:48 p.m.

The Lance Armstrong and Manti Te’o scandals exposed some deep flaws in the professionalism of sports journalists: a willingness to believe the unbelievable and not dig deep lest we scratch the shiny surface of a perfect story.

Both made headlines — and I even wrote a column on the subject — but there is perhaps a more subtle and far more pervasive problem in the world of sports journalism: the (ab)use of statistics.

First things first, numbers are an invaluable tool to any sports journalist; they allow writers to back up arguments with hard incontrovertible facts. But perhaps too often they distract us from the truth; sometimes even the unjustifiable can appear justified through statistics.

Michael Lewis’ bestseller “Moneyball” explained pretty effectively how a cold, hard approach to statistics was behind the Oakland Athletics’ improbable ability to make the playoffs year after year, arguing that the rest of the MLB’s unwillingness to really embrace sabermetrics presented a huge opportunity for the A’s.

But much of that assumes we start with the right statistics in the first place. Even then, baseball is a very different beast compared to most sports. Play is broken down so much that it becomes possible to win games using a deeply mathematical analysis of skills. That is rarely feasible in other sports. Perhaps in the tactical nature of football, but even then, the game can break down and players find themselves forced into a role that none of their personal statistics can have prepared them for.

There are simply too many things we can’t measure because they are too subtle or because we just would be overwhelmed with information. Statistically, junior forward Chiney Ogwumike is by far Stanford women’s basketball’s best player this season. Without her almost guaranteed double-double performance in every game, the Cardinal might not have a 22-2 record so far this season. In comparison, a player like junior guard Sara James may have had a statistically solid season, but has not exactly lit up the box scores.

Except that ignores all the intangibles of James’ contribution. Few players look as fired up as she often has this year, diving all over the court to force jump balls, spurring her team on and providing that spark of determination that can turn the momentum back in its favor. None of those show up in the numerical story of a game, but without a doubt they have had their effect on the scoreboard.

Last week, the BBC released a league table of managers in English soccer. As a patriotic subject of Her Majesty, it pains me to say it, but if you are looking for an example of journalistic statistical failure, this surely was it. The BBC did nothing more than add up the number of points each manager had won — three for each win, one for each tie — and divide it by the number of games they had been in charge.

Perhaps unsurprisingly, Manchester United’s Sir Alex Ferguson — who has won countless trophies in his time at Old Trafford and whose team is now 12 points clear at the top of the Premier League — came in first, but below him the rankings quickly became nonsensical.

For example, Doncaster Rovers’ Brian Flynn was sixth, out of a total of 122 managers. No offense to Donny, which currently sits second in League One, but there is a huge difference between competing in the top flight of English soccer — arguably the best league in the world — and fighting it out two tiers below. But even that isn’t the most embarrassing error of judgment made by the BBC. At the time that table was released, Flynn had been in charge for just three games. While his 2.0 points per game might look impressive, you don’t need to have taken even a single statistics class to realize how meaningless it is to draw conclusions after so few matches.

On this basis, a new manager drafted in for just a single game could become either the best ever coach in the history of sport or the absolute worst, all depending on that one outcome.

Perhaps I’m taking this a little too personally — after all, Reading FC’s Brian McDermott, the Premier League’s Manager of the Month for January was ranked just 99th — but I really don’t think it is unreasonable to have high expectations of major news organizations.

In the Internet age, a simple relaying of statistics — even with a little bit of mathematical manipulation to squeeze out some new numbers — is no longer a job for real journalists. Live box scores already provide fans with the same resources that many of us use to study and analyze games. If we really want to add value to a story, journalists should be reading between those figures and, in order to tell the real story, sometimes even ignoring them entirely.

Tom Taylor’s most-read column of the last 30 days comes in at 239th place in total page views. To console Tom on the “small sample size” of a 30-day window, email him at [email protected] and follow him on Twitter @DailyTomTaylor.



Login or create an account