An important part of the scientific process is not only documenting your successes but also documenting your failures. This is an exercise in the latter.
Hypothesis: There is no such thing as a good crosser.
With the surging influence of attractive, possession-based soccer across the world being spread through the medium of certain "tiki-taka" teams like FC Barcelona, sprinting down the flank and swinging in a hard cross has grown a bit old-hat. Legacy, even.
For the 2012 MLS season, Opta's Chalkboards suggest that crosses were only completed around 26.7 percent of the time. In modern soccer, with world-class midfielders boasting pass completion percentages in the 90th percentile, why would a team ever risk whipping in a cross?
And, even when a "good" cross is attempted, by my eye, there seemed to remain significant variance regarding whether or not it would be completed successfully. There are players across all levels that misplace even the most simple 10-yard passes – how can anyone hit swerving 40-yard balls with similar precision?
So, leaning on a statistical concept called null hypothesis testing, I set off to determine if we can reliably segment good-crossing MLS players from poor-crossing players.
To test this, I split all MLS players into two groups: players that cross the ball frequently and players that do not. The frequency threshold for this segmentation was based on the the ratio of crosses to attempted passes, with players who attempt at least 1 cross for every 10 attempted passes defined as a frequent crosser. Out of 402 sample-size-eligible MLS players, 37 crossed the ball at or exceeding this rate.
My theory here, and it's a bit of supposition, was that players who generally cross the ball at a higher rate do so for a reason. Reasons for this range from a mental predisposition that they are good crossers to simply being presented with crossing situations at a high frequency.
When we compare the crossing statistics across these two groups, it couldn't be more clear that I was wrong. In fact, I'm entirely confident that there is a statistically significant difference between the two groups.
|Frequent Crosser||Completion Rate||Crosses||Completed Crosses||Total Players|
This selection of MLS players is 10.8 percent more accurate than the general population. And, curiously, these 37 players (nine percent of the player pool) accounted for 33.2 percent of MLS' recorded crosses in 2012.
Clearly, "good crossers" exist, and their talents are recognized and exploited by their clubs far more than I expected.
For the sake of completeness, it's important to recognize some of the factors that could have biased the work done here. Some teams are naturally geared to generate crosses from certain players.
For example, when Brad Davis is crossing the ball (14 percent of Davis's passes are crosses), the game is going according to Houston's plan. When Will Bruin is crossing the ball (1 percent of Bruin's passes are crosses), Houston's got a problem.