Roman Burki’s tenure as Borussia Dortmund’s head goalkeeper has been contentious. The Swiss goalkeeper has received plenty of criticism, and has struggled to win over a significant chunk of the fan base. Whenever BVB concede, you can be sure that someone, somewhere is blaming Roman Burki. Even if Marwin Hitz is in goal. But it’s not as though Burki helps himself. He’s a constant source of frustration, even among his fans.
There’s no doubting there are the makings of a good goalkeeper there, but he always seems to fall short of expectations. At his best, Burki is an agile, athletic goalkeeper who possesses good reaction speed, is comfortable with the ball at his feet, and is quick across the ground. However, he is wildly inconsistent, prone to some incredible howlers, and his presence between the sticks seems to create a level of chaos that any team could do without, but especially not a team that is already living and breathing chaos. Simply put, while Burki has the capacity for brilliance, he does not inspire confidence.
Burki’s best season with Borussia Dortmund was probably the 2018/19 season, Favre’s first season in charge of the club. Throughout that season his assured presence in goal helped push BVB to within a point of Bayern Munich and just shy of the Bundesliga title. He was still prone to a stupid moment or two, but his overall performances were of a much higher standard than we have seen since. This season, his performances have raised questions about his future at the club, despite having only recently signed a contract extension. So the question remains: Is Burki good enough to be BVB’s starting goalkeeper? And if not, who is?
Over the next week, I will take attempt to answer these questions, and provide some clarity to a discussion that has been ongoing among the fan-base ever since Burki joined the club. The analysis will be split into four parts, published as four separate articles. Parts one and two will focus on the last four seasons, the seasons for which the advanced metrics data are available, comparing Burki with goalkeepers in the big five leagues. Part One will focus on Burki’s shot-stopping, while Part Two will look at other aspects of a goalkeeper’s duties. Following this, I will take a closer look at performances so far this season, and what this means for Burki and Borussia Dortmund. Finally, Part Four in the series will consider realistic alternatives that BVB could pursue in the summer.
Beyond Shot-Stopping: Analyzing Goalkeepers
To help make this series more accessible, I will start by providing a little information about the way that I will analyze goalkeeper performance, and a short glossary to help define some key terms. The analysis will focus on three areas that broadly define a goalkeeper’s role: shot-stopping, shot-prevention, and distribution. The following is the structure of the analysis, and the primary measures (all data taken from FB Ref) that I will use for each area of a goalkeeper’s role:
- Post-Shot Expected Goals (PSxG)
- Defensive Actions Outside the Penalty Area (#OPA)
- % Crosses Stopped
- Passes Attempted & Completed
- Long-Ball Passes (40+ yards)
- Progressive Pass Distance
I will measure shot-stopping using post-shot expected goals (PSxG). While goals allowed and save percentage have historically been the way that goalkeepers have been judged, they do a poor job of accounting for the quality of shots a keeper faces. PSxG is a metric developed by StatsBomb that measures the probability of a goal after the shot has been taken, as a means of evaluating the quality of shots a goalkeeper faces, and therefore approximately measuring their performance using this as a baseline. It relies on information after the shot has been taken, up until the point that the shot passes the goalkeeper, such as the shot’s trajectory, speed, and the y, z location at which it is estimated that the shot will enter the goal mouth. PSxG filters out shots that are not relevant to the goalkeeper, so only factors in shots that are saved or scored, and it does not include goalkeeper position, as this is a factor in shot-stopping performance. If you’re interested, I’d highly recommend this StatsBomb article introducing PSxG.
Shot-stopping is the marker by which any goalkeeper is traditionally measured, and any world-class keeper has to be a net-positive shot-stopper. But there’s a lot more to a goalkeeper’s role. In a dominant team, a goalkeeper’s duties often lie beyond their own penalty area, as goalkeepers increasingly act as sweepers, coming out of their area to handle threats that occur in behind their defense. This is measured using a goalkeeper’s defensive actions outside the penalty area (#OPA). A goalkeeper that is effective at coming off their line and snuffing out opposition threats outside their box allows their defense to push higher up the pitch without having to worry about the threat posed by an opposition forward getting in behind them and wreaking havoc. Manuel Neuer’s role for both Bayern Munich and Germany over the years is a good example of this, as he has allowed both teams to play with much higher lines, compacting the space for their opponent’s to play, and putting more bodies in threatening positions in the opponent’s half.
On the other hand, goalkeepers also have to rule over their own penalty area with an iron fist. They have to be able to intercept crosses into their box, reducing the opponent’s ability to create offense by aiming for the big man up top. If a goalkeeper is effective at preventing aerial threats, this further limits the means by which an opponent can get shots on goal. This is measured using the percentage of crosses into the penalty area that the keeper stopped (% Crosses Stopped).
Finally, distribution represents a goalkeeper’s contribution to a team’s offense. A goalkeeper can contribute to a team’s possession retention and recycling, as well as progressing the ball up the field, and springing quick counters or pinning the opponent back in their own half. While we typically think of goalkeepers as defensive players, the very best keepers are also offensive weapons, especially in dominant teams. This definitely applies to Borussia Dortmund, who should be looking to their goalkeeper to contribute to the buildup effectively and should be comfortable being an outlet for possession retention. I will measure goalkeeper distribution using the passes attempted and completed, including long-ball passes, which are those passes longer than 40 yards, and finally progressive pass distance, measured in total yards.
So I’ve introduced the question, and provided a rundown on how I plan to analyze goalkeeper performance. If you, like many BVB fans, have strong views about Roman Burki’s performances for Borussia Dortmund, then hopefully this series will be of real interest to you.
This series has been in the works for a little over a year now, so the hope is that it doesn’t simply tell us what we already know, but can also help to shine a light on Burki in a way that says something new.
If you would like to try out some of the code used in this series yourself, the GitHub repository is available here.
I’d also like to extend my thanks to everyone that has helped in shaping this project. That includes my brother, Owen, the goalkeeper in the family, and the members of the FTW Discord, especially fellow writer Steve Zimmerman, who have given me plenty of feedback and discussed Burki’s performances over the last year.