Statistically Speaking

Soccer, more than any of our major American sports, is a subjective game. The lack of scoring opportunities limits the numbers of stats that can be kept, and without stats it can be difficult to judge player performance.

What's the equation for an overhead kick?

Soccer telecasts flash Chyroned graphics with information such as shots, passes completed, and distance ran, but independent of context those stats mean little, like judging a pitcher based solely on his ball-strike ratio. Where were those passes going? Was he running so much because he was always two steps behind the play? Were his shots weak piddlers straight into the keeper’s arms or surface-to-air missiles that end up in the twentieth row, and which would be better? [1]
Yes, the guy who scores lots of goals and gets a lot of assists is a valuable player — Lionel Messi’s dominance can be proven statistically — but once you get below the level of super- and plain old regular stars it’s more difficult. How many goals does one need to score in a season to be a good player? How many assists must one serve up? And is that enough? The soccer press speak derogatorily of strikers who “only score goals,” those players who either get their name on the score sheet or else they contribute nothing to the team. Similarly, both the U.S. and England started strikers at last year’s World Cup who rarely score goals for their country, yet were assumed to be important members of the team.

In his essay on the disappointment inherent in the sports memoir genre, “How Tracy Austin Broke My Heart,” David Foster Wallace writes that the stories of athletes appeal to us “because of our twin compulsions with competitive superiority and hard data.” Soccer lacks hard data. The most common complaint of the soccer doubter — “but they’re not doing anything,” really means, “But they’re not doing anything that can be quantified.”

The former is true of baseball, where there are somewhere between 55 and 500 distinct moments of gameplay, each lasting between 0.396 and 30 seconds.[2]

This home run trot took 30.59 seconds. No, it's not really gameplay, but we're counting it to make a point.

The rest — the jogging in and out of the dugout, the receiving of signals, the scratching oneself in the batter’s box — is fluff.

The average three-hour football telecast, a 2010 Wall Street Journal article reports, consists of 11 minutes of actual gameplay, 17 minutes of replays, and 152 minutes of shots of coaches, players walking back to the huddle, players standing on the sidelines, famous people standing on the sidelines, owners sitting in their boxes, owners standing on the sidelines,[3] cheerleaders, attractive female fans, unattractive male fans, and commercials.

Soccer, by contrast, provides 90 – (TAS[4] – STA[5]) minutes of gameplay. So really, they’re doing far more than the teams in other sports,[6] but those eleven minutes and <500 gameplay events each provide a myriad of measurables to be digested. That which cannot be codified into the mounds of data, the intangibles such as a receiver running a route to force the safety to hesitate and allow his teammate to get open, or a cornerback slipping in coverage and still catching up to his man in time to defend the pass — can be rewound and replayed ad infinitum as soon as the play is over, something soccer broadcasts don’t have time to do, as they’re still showing live action.

Okay, so it's not like soccer broadcasts are innocent of the "attractive female fan" distraction.

No other sport’s journalists make common practice of providing player ratings as part of their postgame coverage. They can let the statistics speak for themselves; how many hits, three-pointers, and interceptions as an indication of how well a particular player did. Player ratings are box scores minus the facts; subjective measures of a player’s performance.

We all know soccer is making progress in this area. Chalkboard technology, with data provided by Opta Sports and made publicly available by the Guardian and a variety of iPhone apps, among others, track stats in context, showing passes, shots, and tackles and their location on the field. You can filter crosses made by a particular player or a team’s interceptions in their own half between the twentieth and fortieth minutes. Shots and passes come with arrows attached showing their direction and rough end point, and are colored to show whether they were successful or not, on target or not, or assists/goals or not.

Brian Phillips tells us that clubs and European companies are keeping their own statistics, with their own algorithms and weights assigned to in-game action, and guard them like they’re Coca-Cola’s 7X. For now, efforts to sort through the data that is available to the public, to tell us what it all means, remain somewhat fishy. Opta also provides data for the Castrol Rankings, a supposedly objective measure of who the world’s best soccer players have been in the last year, based on where a player touches the ball and what they do with it in club competitions in the world’s five biggest leagues.[7] Nate Silver, the statistician who became famous after switching from crunching baseball statistics to crunching political polling data, put together the Soccer Power Index for ESPN, which uses a complex algorithm involving goal differential, level of competition, and a combination of game-and player-rating datasets[8] to predict which national soccer teams are currently the world’s best.


Luis Amaral, an engineering professor at Northwestern, published a paper after the 2008 Euro tournament in which he and his team analyzed each of the 16 competitors’ “networks,” how they interact with each other and the ball, basically who is passing to whom and what comes of it. Amaral describes his work in a 2010 Forbes article as trying to figure out “who is responsible for what.” Spain, the tournament’s winner, and Xavi, its MVP, finished with the highest scores.

But of course, we already knew that, didn’t we?


[1] The answer is the first one — just ask Robert Green, who I realize I’m making fun of in two straight articles — but watching a game you’re more likely to forgive someone who hits the latter kind than the former.


[2] Assuming a 103 MPH fastball and a really slow home run trot, which we’re generously counting as gameplay.

[3] That is, Jerry Jones.


[4] Time actually stopped.


[5] Stoppage time allotted.


[6] Basketball excluded. Since a basketball game lasts only half as long as a soccer game, there’s an argument to be made there, but I think those of us who are reasonable can agree that it’s something of a silly one.


[7] Their metrics, however, are either impossibly complex or really stupid, because as of this writing Karim Benzema is their second best player in the world, and Maxi Rodriguez is two spots above Xavi. Also, Castrol has signed Cristiano Ronaldo as the spokesman for the Castrol Rankings, which to my mind casts something of a pall over the objectivity of the endeavor, even if Ronaldo, who as of June 2011 is third in the rankings, is still somehow undervalued by them.


[8] According to its FAQ on ESPN, the weight given to a national team’s performance in international games and its player’s performances in club games differs from team to team, which, to make a long story short, almost certainly ends up in England being overrated.

This entry was posted in Culture and tagged , , , , , , , . Bookmark the permalink.

2 Responses to Statistically Speaking

  1. Pingback: These Magic Moments | The Other 87

  2. Pingback: Our Greatest Hits: These Magic Moments | The Other 87

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s