So far, over the course of this series analyzing the Hinrunde, the results have suggested that Bayern and RB Leipzig are the best teams in the league this season, while Favre’s conservatism is a perfect accompaniment to Dortmund’s propensity for errors, and those errors have been a consistent problem that has contributed significantly to Dortmund’s underperformance in recent years.

That seems awfully bleak. So maybe I can find some positives in the final article of this series, by taking a look at some of the best individual performances in the Bundesliga this season.

The League’s Best Offensive Players

In terms of individual production, the obvious place to start is in terms of goals. We know that Robert Lewandowski and Timo Werner are leading the league with 19 and 17 goals respectively, while Reus and Sancho have been Dortmund’s most potent goalscorers so far this season with 9 goals each. Sancho has also been Dortmund’s best creator, with 9 assists, just behind Thomas Muller who has so far created 10 goals this season. Thorgan Hazard is just behind these two, in joint third place with RB Leipzig’s Christopher Nkunku, both of whom have 7 assists. But how about in terms of expected production?

At the very top of both the xG and xA tables things are looking much the same. Lewandowski has produced the highest xG totals, while Muller also leads the xA table, but analyzing non-penalty expected goals per 90 (NPxG/90) and expected assists per 90 (xA/90) suggests that Werner and Gnabry have been the standout performers. For Dortmund, Paco leads the NPxG/90 rankings, with 0.63. Reus and Sancho are a little behind him with 0.46 and 0.3. In terms of xA/90, Sancho tops the table with 0.29, while Hakimi and Hazard are not far behind with 0.23 and 0.22.

Figure 1 presents the results. Shout out to Ryo Nakagawara for the inspiration for this plot and provision of some very clean, reproducible code as a base to build from.

Werner & Gnabry have both produced well above average NPxG/90 and xA/90, and are two of the only five players that are in the 95^th percentile in both categories, the others being Hermann, Plea, and Nkunku. Though Dortmund don’t have any clear standout performers in Figure 1, Sancho, Reus and Alcacer all feature.

Dortmund’s Passing Numbers

Given that Sancho is leading the assists and the expected assists for Dortmund, is he the primary outlet for chance creation? I compared the chance creation in terms of first and second assists to answer this question. Second assists are the pass that leads to the assist. Thanks to Adam Darowski for compiling the second assist data, and for sharing it with me to make my analysis a little easier! I highly recommend taking a look at his site, and his article that goes into a little more detail on the second assist data. I use his “hockey assists” definition to count second assists in this article.

Figure 2 plots Dortmund’s primary creators, split by first assists and second assists.

The results show that chance creation is split across a number of sources. Though Sancho is the leader of the pack, Brandt, Hazard, and Hakimi are all chipping in with their fair share of creativity. It’s especially interesting that Sancho, Hazard, and Hakimi have all produced more first assists than second assists, while the opposite is the case for Brandt, who leads the table in terms of second assists, with an impressive 8. Brandt has been fantastic so far this season, acting as a deep facilitator behind the likes of Sancho, Reus, and Hazard. However, he loses marks for this atrocity.

When the numbers are combined, they really demonstrate how absurd Sancho’s production has been so far this season. Figure 3 demonstrates this case.

Sancho has created a total of 17 chances, while Hazard has 10, and both Brandt and Hakimi are on 9. Sancho has been having a monster season, and with each game he adds another few million to his transfer value. At this rate we might be able to buy a whole new squad with the proceeds in the summer.

Splitting Sancho’s performances by game only serves to further demonstrate the incredible production he brings to the table, as Figure 4 shows.

There has been only 5 games this season that he hasn’t chipped in with at least one goal or assist, and when second assists are included, he has had two monster games with 5 goal totals. It’s interesting to note that his xG & xA numbers are not at the same level, but he’s been able to maintain this level of overperformance throughout his time at the club, so it’s possible that this is just the level of player he is. He’s a remarkable talent, that’s for sure.

To further illustrate the impact that Brandt and Sancho have on Dortmund’s attack, they lead the team in terms of passes into the penalty area. Digging deeper into the passing numbers also demonstrates how well Dortmund’s attacking responsibilities are spread across the lineup. As well as Brandt and Sancho, Hakimi, Hazard, Guerreiro, and Reus are all heavily involved. In terms of the buildup play, the duties are shared by Brandt, Hummels, Akanji and Witsel. Witsel has been one of Dortmund’s best players this season, and even boasts the highest pass completion rate in the Bundesliga.

On the flip side, the errors are pretty evenly spread around the squad too. Hakimi has been the main culprit, having committed two errors so far this season, but Brandt, Witsel, Akanji, Burki, Hummels, and Piszczek have all contributed.

Comparing the League’s Best Players Using On/Off & Plus/Minus Metrics

Comparing individual performances is a difficult task because it is often impacted by the level of the team performance around the individual too. It’s no surprise that Bayern and Leipzig top the table in terms of individual performances, because they do so at a team-level as well. But it’s possible to compare team performances with and without each player, in an attempt to test the impact that they have had on their team when they are playing. Adam’s discussion of these metrics is also well worth a look.

Figure 5 plots the impact that individual players have had on their teams, both in terms of expected and actual goals when they’re on and off the pitch.

Sancho and Hazard are Dortmund’s standout performers. Interestingly, Sancho’s xG & xA numbers may not be that staggering, but he has one of the biggest impacts on team xG of any player in the league.

It is a little tougher to measure the worst players, primarily because it is easy to give players the benefit of the doubt that their good performances are their own doing. Nonetheless, Figure 6 plots the reverse of Figure 5.

As demonstrated by Mats Hummels, On/Off metrics don’t always capture the full story when comparing football players. The best players will spend very little time out of the game, making it difficult to draw conclusions about players that are playing the majority of their teams’ minutes. Hummels was on the pitch for the full 90 minutes of the thumping Dortmund received from Bayern, but he didn’t play in the 5-0 win against Dusseldorf, due to suspension. Neither of these results happened because of Hummels playing or not, but they have a significant impact on his On/Off numbers. That is why these numbers have to be taken with a pinch of salt and a dose of context.

Turning to xG Plus/Minus, I have plotted the performances of Dortmund’s squad so far this season, as shown in Figure 7.

In line with the rest of the findings so far in this article, Jadon Sancho has been Dortmund’s best player this season. A little behind him, but still performing very well, are Reus and Witsel. There are only five players that have negative xG while on the pitch, and of those only Weigl and Schulz have played significant minutes. These numbers don’t reflect well on Nico Schulz, who has been the worst of the summer’s signings.

The Bundesliga’s Best Goalkeepers: Modern Vs Traditional Goalkeeping

Turning to performances at the other end of the pitch, I compared Burki with the rest of the Bundesliga’s keepers. Figure 8 plots the best shot stoppers in the league this season, based on their shots saved above average (SSAA) and their on target expected goals plus/minus (oTxG +/-).

Burki actually performs quite poorly according to these metrics. While his oTxG +/- is relatively average, it is still negative, and when coupled with his particularly poor SSAA, gis performances start to look quite bad. Despite being a big fan of Burki generally, I do wonder if his occasional moments of brilliance are papering over certain cracks. At times his defense doesn’t do him any favors, but that doesn’t appear to account for his poor numbers entirely.

While I strongly believe in using data analysis as an accompaniment rather than a replacement for the “eye-test”, it is also necessary to guard against bias. If there is no obvious explanation for the data and the eye not matching up, that may be a sign that the eye is wrong. Perhaps Burki has been sub-par this season?

However, it is also worth noting that modern goalkeepers are not glued to their line, and are increasingly required to be comfortable with the ball at their feet. Modern goalkeepers are more involved in buildup play, and they frequently contribute defensively beyond their own penalty. Figure 9 demonstrates this.

Neuer, Burki, and Gulasci all look like the archetypal modern goalkeeper, frequently being involved in defensive actions outside their own penalty area. On the other hand, Sommer, Nubel, and Hradecky, three of the league’s best in terms of shot stopping, are all more reserved, and more likely to stay close to their penalty area.

However, it is also worth noting that by being so active off their lines, the likes of Neuer and Burki may cut out opposition attacks that the other goalkeepers don’t. Their willingness to come out of their penalty area and defend also allows their defenses to play higher lines without having to worry about the quick counters as much. On top of this, Neuer has an incredible pass completion rate of 90.4%, while also ranking in the top 5 for passes attempted per 90. Burki is less impressive in this regard, simply because he attempts fewer passes, but he still maintains an impressive pass completion rate (79.3%). Finally, while Sommer doesn’t make many plays outside of his own penalty area, he does perform very well with the ball at his feet. He ranks 3^rd in the Bundesliga for both pass attempts per 90 (45.18) and pass completion rate (81.4%). At this point, I think Sommer is the best goalkeeper in Germany. Maybe Dortmund should think about replacing one Swiss keeper with another?

Conclusion

So there were some positives to take from the Hinrunde. Jadon Sancho is a freak of nature and Julian Brandt is an artist that chose to express himself with a ball at his feet. It is also impressive to see how much the Dortmund attack shares responsibility for creating and converting chances. Finally, individual performances by Hazard, Hakimi, and Reus should give Dortmund fans a little reason to feel excited about the second half of this season.

To finish the Hinrunde Review, I’d just like to thank everyone that contributed with their feedback, data, code, and their bleeding eyes as they soldiered on through these stats-heavy, several thousand word behemoths. I’ve enjoyed the discussions we’ve had on each topic, and it has opened my mind to various different theories and ideas that I hadn’t considered before. Hopefully the second half of the season will give us a little more to cheer about!