Everyone says Borussia Dortmund are in crisis. Less than two years removed from making the Champions League final and less than three years removed from winning the Bundesliga, BVB have fallen into the relegation zone. With the league resuming this weekend after its annual winter break, Dortmund sit 17th, a full 30 points behind first-place Bayern Munich.
Proposed reasons for Dortmund's slide have ranged from injuries to lack of belief to the league catching on to Jurgen Klopp's tricks. Other clubs, including Hoffenheim and Bayer Leverkusen, have developed their own versions of his high-pressing, transition-oriented attack, and Klopp's side has not been able to overrun opponents as they used to.
Yet statistical analysis offers another explanation. For a start, the crisis at Dortmund is not as serious as it first seems. Dortmund have been struggling to finish their chances but they are still creating quality scoring opportunities at a good rate. Shooting slumps like this are reasonably common even at the highest level in soccer, and they rarely last.
In soccer, the rate at which shots are scored is massively variable. A club might win 1-0 on a perfect long-distance strike one week and then lose 2-1 the next while firing a dozen shots from good positions right into the keeper's chest. One of the counterintuitive lessons of statistical analysis in soccer is that the latter team, the one that lost 2-1, generally has better odds to win future matches.
This season, Dortmund have been the very model of that "losing" team. They have attempted 288 non-penalty shots and conceded just 158. The only other two clubs in the Bundesliga with 100 more shots attempted than conceded: third-place Leverkusen and league leaders Bayern Munich. Not all shots are equally valuable, of course, but BVB's numbers still look good under closer analysis.
Of BVB's shots, 109 have been attempted from the "danger zone," the region in the center of the 18-yard box from which most goals are scored. This total is third most in the Bundesliga, behind only Bayern and Wolfsburg. But Bayern have scored 41 goals and Wolfsburg 33, compared to just 16 for Dortmund. Again and again, Klopp's side has created a better set of shooting opportunities than their opponents and come away without any points to show for it.
To estimate the quality of chances a team creates, I am using a statistic called "expected goals." This statistic gauges the likelihood of scoring for each shooting chance based on location, the type of pass which assisted the shot, which body part is used to attempt the shot, the speed of the attacking move leading to the shot and several other factors. (You can read the method in full here.)
According to expected goals, a typical team should score 25 goals from the chances Dortmund created, but BVB have netted just 18. On the other side, Dortmund's opponents have scored 26 goals on chances that would ordinarily result in about 17. So while Klopp's side has a real goal difference of minus-eight, its expected goals difference is plus-eight.
There are three clubs in the Bundesliga with goal difference of between plus-seven and plus-nine: Leverkusen, Borussia Monchengladbach and Schalke. They stand in third, fourth and fifth place respectively. Yet failure to convert chances has taken Dortmund from Champions League contention (where they should be) all the way into the relegation zone.
To illustrate this point, the following graphic displays the locations of the 30 best chances that Dortmund has created or conceded this season. Black markers represent goals, yellow markers represent shots that were not scored. The size of the marker represents the estimated expected goals value of the chance.
On the left-hand side, those clusters of big yellow boxes in the center of the penalty area represent good chances that Dortmund's attackers couldn't finish. The smaller black boxes on the right show the locations of tougher chances that Dortmund's opponents knocked home.
What does this mean for Dortmund? Is there something else wrong with the club that expected goals fails to recognize? To check, I looked at the team's performance in the Champions League. Dortmund has a plus-10 goal difference in group play, compared to about plus-seven expected goals difference. Outside of the Bundesliga, Klopp's side has had no trouble converting chances.
Next, I identified clubs over the past five seasons in the biggest leagues in Europe that similarly struggled to convert chances early in the season. Did these clubs' shooting woes continue? On average, they did not.
I found 27 clubs that had underplayed xG by at least nine goals. To put the numbers on the same scale, I am using the statistic "goals ratio" which is calculated as goals scored divided by the sum of goals scored and goals conceded. These clubs were outscored about 2-to-1 over the first 17 matches of the season but managed roughly even goal difference statistics over the remainder of the season.
It is expected goals ratio, not actual goals ratio, that best predicted how these teams would close out the season. Statistical analysis therefore predicts a Dortmund turnaround and a strong run of results to come.
These teams include several of the most impressive second-half turnarounds in recent years but only six had an expected goals difference as strong as Dortmund's this year. Their identities make for heartening reading if you're a BVB fan. There's Liverpool in 2012-13, who began the year with just a plus-one goal difference but a nearly plus-11 expected goals difference. In the second half, the Reds outscored their opponents by 23 goals and the very next year they contended for the title. The other five are VfB Stuttgart and Chievo Verona from 2009-10, Everton and Atletico Madrid from 2011-12 and Tottenham Hotspur last season.
For Bundesliga fans, the most notable team comparison is VfB Stuttgart in 2009-10. After finishing third the year before and qualifying for the Champions League, Stuttgart started their 2009 season in a terrible slump and were in 15th place at the winter break, only sitting above the relegation places on goal difference. The club had been outscored by seven goals despite an expected goals difference of plus-five.
Yet Stuttgart was a juggernaut from that point on, taking 39 points from its remaining 17 matches and posting a stellar plus-17 goal difference. The club did make some midseason changes, most notably sacking manager Markus Babbel, but statistics suggest that Stuttgart's comeback was always in the cards as their attackers started to convert good chances they created.
There are further parallels to Dortmund's current plight. Stuttgart lost striker Mario Gomez to Bayern Munich in the summer of 2009, just as Dortmund lost Robert Lewandowski to the Bavarian giants in the recent summer transfer window.
Of course, no club with four wins from 17 games is without flaws. There are significant issues in defense and even if BVB's plus-eight expected goals difference correctly estimates the club's quality, that is a far cry from last season's plus-42. A recent study by Colin Trainor at Stats Bomb found that Dortmund's attackers do not seem to be contributing as many high-quality chances as they did last season.
Dortmund has indeed got worse and will struggle to make the Champions League again. But do not be surprised if February or March sees a spate of "Klopp is back" articles, as Dortmund race back into contention. The club's underlying stats remain strong and project a bright future.