Last season, D.C. United had one of the worst attacking seasons in Major League Soccer history. The black and red scored just 22 goals and allowed their opponents to keep a clean sheet 16 times. This year, things have changed. United has scored 16 goals in 11 matches, rebounding to third place in the Eastern Conference. It may be too early to start printing playoff tickets, but it is not too early to say this has been a remarkable turnaround.
Peculiarly, even though D.C. United is scoring more, the club is taking roughly the same number of shots per match that it did last year. D.C. United averaged 11.7 shots per match last season compared to 11.8 this season. What has changed, however, is the locations from which shots are attempted. In my last article, I introduced the concept of the danger zone, the central area of the 18-yard-box from which most goals are scored. D.C. United may not be shooting more overall, but its players are attempting far more shots from high-expectation areas of the pitch.
Part of the story here is that D.C. United’s attack is doing better at getting into dangerous locations for shots. Better shots lead to more goals. The other part of the story is shot conversion. Last year, United could not finish on goal to save its life, but this season United is finishing at quite good rates. Of 130 danger-zone shots attempted in 2013, D.C. United scored just nine, a finishing rate under 7 percent. This season the club is 10 for 51, a much more typical rate.
To quantify the quality of chances a team creates, I use the statistic “expected goals.” This stat estimates the likelihood of a shot being scored based on a number of variables. Location on the pitch is the most important, but xG also considers the type of pass that assisted the shot, whether the shot was a header and whether it was taken from a direct free kick. D.C. United has a better rate of expected goals this season than last, but the difference is much less stark than the difference in actual goals.
So what is going on here? D.C. United completely overhauled its roster in the offseason. Is it just that these new players are better finishers and thus scoring more goals?
This is a common-sense explanation, but it is statistically dubious. Rates of shot conversion in soccer have a high degree of randomness. Goals scored is a weaker predictor of future goals scored than either total shots attempted or expected goals scored. At the same time, conversion rates are not merely a function of “luck.” Better players do score at higher rates. As an illustration, I have taken a look at rates of goals scored per attempt from the danger zone in MLS. I have broken down the 2012 and 2013 seasons into two halves, and compared the rate of danger-zone shot conversion by a club in the first half of the season to its conversion rate in the second half. As you can see, there is almost no relationship.
For almost all MLS clubs in the last two years, danger-zone shot conversion has been highly variable. The exception, of course, is United in 2013. It was just as bad at finishing its chances in the second half of the season as it was in the first half. United probably was not quite as bad as its goals-scored total suggests. But it is possible that United was the rare club that just could not finish. In either case, with a new roster and much improved underlying stats, the horrors of 2013 are unlikely to be repeated this year.
Data supplied by Opta unless otherwise noted