Opposition analysis: Bayern Munich
How Arsenal can beat Bayern. Detailed statistical analysis, strengths to nullify, vulnerabilities to exploit, learnings from the tape, and much more
“We have achieved nothing at all as yet; we have not made up our minds how we stand with the past; we only philosophise, complain of boredom, or drink vodka. It is so plain that, before we can live in the present, we must first redeem the past, and have done with it; and it is only by suffering that we can redeem it, only by strenuous, unremitting toil. If you have the household keys, throw them in the well and go away. Be free, be free as the wind. Believe what I say, Anya; believe what I say. I'm not thirty yet; I am still young, still a student; but what I have been through! I am hungry as the winter; I am sick, anxious, poor as a beggar. Fate has tossed me hither and thither; I have been everywhere, everywhere. But wherever I have been, every minute, day and night, my soul has been full of mysterious anticipations. I feel the approach of happiness, Anya; I see it coming.”
— Anton Chekhov, The Cherry Orchard
Do you forgive me these overserious quotes?
It’s only football, after all.
Oh well.
Today is a big day, as I needn’t tell you. The schedule is coming fast and thick, and it is hard for your moonlighting writer to keep up with it all. Since we last spoke, Manchester City were fought to a draw. Luton were vanquished with a heavily-rotated lineup. Arsenal played Brighton at the Amex — a fixture I was frankly nervous about, the more I dug in — and it resulted in a 3-0 drubbing. Liverpool felt the flipside of variance. Arsenal are in first place and in full control of their fate.
Now it’s back to the Champions League.
More precisely, it’s back to a familiar foe.
“My soul is full of mysterious anticipations.”
📃 Our opposition report
Last time out, we introduced something new we’ve been working on: a tool for researching opponents. The idea was to quickly surface novel insights that may bring focus to an upcoming matchup. If you missed it, here’s the skinny.
To oversimplify, I’d wanted to do some correlation studies to see which statistics — be they individual or team-level — had the greatest relationship with the performances of specific opponents.
For instance, if you learn that disrupting a certain kind of pass, or restricting a certain player’s carry, is disproportionately pivotal to a team’s performance, then you can use that to focus your film study — and, ultimately, your gameplan.
I’d built some straightforward measures myself, and I was interested in taking it to the next level. To help do that, I’ve been working with t — a statistician and brilliant mind on such matters. (He’s a Director at Kadre, a data science/machine learning consultancy, and is “still romantically waiting for Abou Diaby to return from injury.”)
We’ve continued with our work, and will keep sharing it as it continues to evolve, even though it’s early in the story.
With that, here’s a “Team Correlations” report for Bayern, with some interesting things highlighted.
I will repeat the caveats of last time:
Correlation is not causation. Just because something happened more frequently, it does not mean it caused the underlying xG difference. “Just stop them from getting corner kicks and prosper!” is not the takeaway here.
Not all correlations are instructive or even important. It’s not all that useful to learn that teams tend to win more games when they shoot a lot. Key passes, assists, touches in the attacking third — these are all likely to come from dominant performances.
There are sample size limitations and general noise, particularly on statistics that don’t happen at a high volume.
Next, here’s an “Areas of Emphasis” graphic, which outlines some of the key things to know about how Bayern play.
There is a ton of data we’re going through — hundreds and hundreds of rows and correlations — so I sorted through and picked five of the most interesting, novel, and (hopefully) actionable metrics for each. This is the stuff that seems to have a disproportionate relationship to a Bayern performance.
And finally, for now, we have our “Player Correlations.” We’ll probably be sharing a huge swath of these correlations in future iterations of this thing, but for now, I’ve grabbed ten of the most interesting things about Bayern, in the form of particular player metrics that may impact (or reflect) their performances the most.
In future editions, we’ll have some expanded views, adjustments based on game-state, some other data sources, and more. We’ll also have these cool decision trees on how to approach certain players. t is smart.
For now, enjoy.
Let’s see what we can parse from it all.
👉 A reintroduction
The most important thing to know about Bayern is that you shouldn’t let the headlines fool you. They are really fucking good. The clownshow at the top of their operation has obfuscated a simple truth, and that is as follows: what happened at the top of the Bundesliga this year is mostly just a story of Bayern being fairly weird while another team decided not to lose any games.
There are only two or three attacking batteries in the world who can chat with the sheer breadth of talent of Kane, Musiala, Coman, Sané, Tel, Gnabry, Müller and the like. In fact, there’s a case that they’re the best, period. They have the single-best goal tally and xG in all of Europe’s top-5 leagues, and the fifth-best xG against. This results in the top expected goal difference in Europe. It’s by a decent margin, too. They lead the way with +1.75 xG difference (per Opta); the second highest is Inter with +1.37, followed by Arsenal at +1.29.
OK, are you properly expectation-managed? Now we can get to their problems.
As I see it, I’d argue that two problems are beleaguering Bayern:
Personnel: We talk about the importance of “complementary profiles” a lot in these parts. Bayern have a surplus of talent but they don’t always pair well together. This is particularly true of the midfield, where any potential pairing is either too slow or too light on tempo-setting passing. This is also true of their CB pairings, which should be a strength on paper; in practice, their coordination has been a bit off, and they’ve often been caught lunging upfield (typically trying to cover space for midfielders), leaving gaps. Rice would have indeed solved a lot of problems for them, but that’s true of almost every club. Sorry!
Tempo: This team naturally comes alive in fluid, direct, bursty moments of attack. In those stanzas, their talent can shine, intuition rules the day, and their ball-striking and dribbling can beat anybody. There are two natural adversaries to this: the first is Tuchel, who can seek to exert patient control in this job assignment (they’ve had more possession than their opponents in 37 of 39 matches so far). The second is just the reality of being Bayern in the Bundesliga: you are going to naturally dominate the ball, whether you want to or not, so it’s going to be harder to manufacture transitional moments.
As you’ll see in this chart, they’re the second-slowest build-up team in the league.
They also have the longest sequence times, with 13.64 seconds per sequence.
…even so, they lead the league in direct attacks, with 70.
Here is a look at some of the lineups they’ve used in big matchups of late. Against Leverkusen, you’ll see his use of a back-three, which could conceivably make a reappearance (even though it didn’t work out then).
Also from Transfermarkt, we can see one of the best views around: a seasonlong squad absences list. With the league out of reach, Tuchel rested many of their stalwarts in advance of our game.
Bayern posted their travelling squad, and it appears the likes of Sané, Neuer, Mazraoui, and Pavlović are on their way back for this one — while Coman will be a later call (though he’s reportedly in training).
Tactically, Tuchel has been tinkering throughout the season, as Tuchel does. We won’t fully know what to expect in this one.
Against Dortmund, we saw Goretzka dropping to the backline to form a flat-three in build-up. Laimer was the lone-6 from there, and Kimmich and Davies kept the width. As a bit of foreshadowing, I’ve highlighted the danger zone when facing them — the central area where players like Kane, Musiala, and Sané like to drop, turn, carry, and playmake.
Otherwise, like against Lazio, we’ve seen a more standard 4-2 build-up structure.
From there, they transition to a more advanced 3-box-3 (or 2-3-5) look that can create a lot of possessional dominance, but not always a genuine cutting-edge.
Out-of-possession, their primary blocking shape has been a 4-4-2, though, again, a 5-back isn’t fully out of the question. They press with medium-to-high intensity (~11 PPDA), and have been effective in the league, generating 40 high turnovers that result in shots.
This, from markstats, can give you a sense of how their responsibilities may morph:
Finally, for this section, I’ll mention two areas:
Set pieces: This is a team strength. They have 15 set play goals (leading the Bundesliga), and the lowest set play xG against (3.57).
Rebounds: They’ve scored a lot of goals on rebounds (11 in league). One must stay switched on, especially when Kane goes a second and third time. The last-line mayhem stuff still feels like an area of growth, and this is a great time to show it (if it can’t be avoided completely).
After rewatching, here’s where I’d focus.
💭 How to beat Bayern
Here are some things I’m thinking of before kickoff. I’ve watched Bayern a fair bit this year, and rewatched three games for this, but blindspots are totally possible, and in truth, I had to get this together quicker than I’d like. Drop any additional context in the comments.
Warning: I’m going to say “duels” a lot.
1. Let them pass it around; deny the dribble
Bayern are not necessarily more dangerous the more of the ball they have. When they control the ball with their slower build-up, they are still proficient on a global scale, but not as dynamic as they could be. Their attackers want space, time, and 1v1s. This is especially stark in the stats:
They have 167 more attempted take-ons than anyone in the Bundesliga, and complete them at the highest percentage.
Statistically, they do disproportionately well (compared to other league teams) when they dribble and carry. To perform at their best, they heavily rely on carries into the penalty area, progressive carries, carry distance, and stuff like that.
Conversely, they don’t do as well with more of the ball. As an example, possession usually has a moderate correlation to underlying performance (0.45 for the Bundesliga); for Bayern, it’s 0.07. Stuff like “attempted short passes” or “live-ball passes overall,” which usually have the expected positive correlation to doing well, are actually in the negative for Bayern.
The preference extends to game states. After they’ve broken the seal, they tend to rain in the goals (2.90 xG/90 when up 1, 3.37 xG/90 when up more than one, per understat). When they go down, and the game stays closed, they are much less effective.
In an even game state, much of the game will depend on pure 1v1 duels.
Kingsley Coman is coming back from injury. If he plays, he’ll be important, as his name kept popping up as impactful to Bayern’s success — and especially when he can beat his man on the dribble. Think of all their disappointing results this year, then consider this anecdotal stat: If Coman beats his man at least once on a take-on, Bayern are 15-2-1.
Bayern also do particularly well with carries from Musiala and Davies, for obvious reasons.
Here’s a look at how Coman can win a situation.
For these reasons, my preference would be to start strong 1v1 defenders, and to push around their dribblers prior to their reception wherever possible, and demonstrate a physical game.
There’s a reason why central dribbling is risky for the attacking team. It can come back at you in a hurry, as Musiala saw here (this resulted in a goal).
Remember when I said it was going to be talking about duels?
2. Bully the middle; restrict access for Kane and Musiala
Here’s a quick view of how Bayern look when they’re at their best. The defensive shape is stretched out, and Musiala drops into the middle to receive. He then turns his man (he’s one of the best in the world at that) and kicks off an attack. From there, great intuition and ball-striking result in a goal. The end of the gif got cut off, so you’ll have to take my word for it.
You can see another example here. The “simple” act of winning this initial duel is often the difference between beating Bayern and them hanging a crooked number on you.
As we’ve outlined, the below yellow area is the place where so many Bayern attacks can originate. The full-backs hold so much width that the defending team has to make a tough decision on whether they can stay compact or otherwise let a player like Davies bomb down the line. When they are stretched, any attacker can drop in here, turn, and it’s off to the races. Kane probably has to drop too low considering the players they have in build-up. But guess what? He makes it work.
A lot of this is about — you guessed it — duels and physicality. Tactically, my preference is to compact the middle of the pitch greatly — honing in on those central receptions — and only try to squeeze the wide full-backs when they’ve actually received it.
This is what Arsenal often do anyway. Here’s the standard mid-block, which shows how crowded it can get once Arsenal drops from their more aggressive pressing shape (which I’d expect in this home leg).
The toughest question (at least for me) is where to place Rice. Contrary to where he might land, my entire impulse is to make him the deepest midfielder so that he can stay close to the likes of Kane and Musiala, and leave the smaller, higher pocket to another player (Havertz, Partey, Jorginho). I’m not sure it’ll happen, and that was my preference against Brighton (to keep Jorginho out of space), and I may have been proven wrong there.
Regardless of where he plays, Rice should offer an upper hand.
In attack, too, the midfield offers the greatest opportunity. Their likely duos can be moved around a bit, and if the moves are crisp enough, they can be left catching up to the play.
(And if you’re wondering, I’m glossing over stuff like “the CBs need to do a good job against Kane” as a given.)
3. Attack the left-back (likely Davies)
Saka should have an advantage on Davies (or Guerreiro). Davies can, for good reason, be overly trusting of his recovery speed, so I’ve seen him turning and lunging at balls. This can make him vulnerable to sleek interplay and 1-2s.
His defending has gone through a few phases in his career — there was a period where I thought it was brilliant — but I’ve seen him losing runners and switching off a few times over the last couple of months. As we saw in the numbers, the more active he is forced to be defensively, the worse Bayern tend to do.
4. Win second balls
Arsenal are an outlier when it comes to second balls.
And this looks like it could be an advantage. With Rice and Havertz offering a pure athletic superiority, Bayern can also not be at their best when there are moments of disarray and handoffs. We saw that against Darmstadt 98:
And just last week against Heidenheim.
The more they unsuccessfully approach aerial balls and duels, the more their defensive order is disrupted, the worse they fare.
5. Make them track runners, switch, and coordinate
Bayern’s biggest problem out-of-possession, as I see it, is the coordination between their centre-backs and their midfielders. The midfielders can get caught up the pitch, the centre-backs often look to compensate, and perhaps because of the revolving door back there, there isn’t the telekinesis we’ve seen develop between Gabriel and Saliba. It’s perhaps starting to steady out with Dier and de Ligt. (I’m generally confused by what they’ve done to my boy Kim Min-jae.)
As a result, movements like the below — in which Havertz drops and pulls Dunk around, and then Arsenal attack the space in behind — are going to be key.
The other place to exploit is on underlaps. Arsenal should test the resolve of Bayern wingers and advanced midfielders to track runners all the way to the byline. If they don’t, make them pay.
🔥 Final thoughts
It’s time to wrap it up so I can get this thing out today.
Fitness questions surround the tie. Here’s what I wrote about the most recent injury report, which was as simple and refreshing as can be imagined from Arsenal’s perspective:
Player fitness is an enormously complex topic, awash with luck, bullshit, and a million little confusing variables. But if you’re able to go into a UCL quarter-final and the injury report is just “every player is out there,” you are probably doing something very right.
As we covered in What Ails Arsenal, written during the depths of the bad run, a lot of work has gone into the squad’s availability at this crucial juncture.
The Premier League is a battle of attrition, and I’d humbly suggest that while many of Arteta’s early-season moves have indeed been borne out of a desire for solidity and control, there may have been another, more practical objective in play as well: the desire to play the long game, better supporting a thinning squad’s fitness for the full campaign, lest the previous season’s travails be repeated.
One way to do that is through steady, advanced possession. By dribbling less and passing around the attacking third more, the opponent has to move around a lot, but you have to run at lower rates. After all, the ball is in transit, not you. This is what I mean when I say that a playing style can affect squad depth and vice versa.
It’s a tremendous achievement.
I can hear you ask me for my XI. Maybe that’s the voice in my head.
From our vantage point, we’re still not working with complete information about the true fitness of particular players — i.e., who is battling a niggle, or who can go the full-90. But considering a few factors: Zinchenko played significant minutes vs Brighton; this isn’t likely to be a full transition spectacle (re: Martinelli, who is working his way back); Jorginho went the full-90, which may point to Partey; and wide 1v1s are at an absolute premium, here’s what I’d suggest is perhaps likely. I’ve put less thought into this than usual.
Again, that’s assuming that Martinelli will be a power-sub, but that’s no given. If he’s ready to do 70’, the decision gets awfully tough. I may be open to Zinchenko for the home leg, but those Coman duel numbers got me weighting that heavily. Similarly, it’s a lot to ask for Kiwior in a 1v1. He rose to the moment against Porto, but he looked how a lot of defenders would look against Bernardo Silva, and is still prone to heavy steps against the quickest wingers. I’ll admit I may have disproportionate faith in his composure these days.
With that said, I do (personally) have this next image in the back of my head, though logic definitely argues against it for this leg. I love the idea of having pure strength and athleticism in the midfield (Rice/Havertz defensive pivot), overrunning them in the duel, playing over the top, and letting them have the ball a bit. This would require Ødegaard as a more full-time second pivot in-possession (with the help of White), and could create a 4-box-2 look that Arsenal employed against Liverpool. It would essentially be double-strikers in possession, as was used during that run where Havertz and Trossard were interchanging and flooding the midfield.
Ultimately, it’s likely too negative and light on passing, though. I could argue that you could just hoof it long in that case, but alas. We can save something like that for the Allianz, though I’d probably swap a striker for a DM and roll with a true double-pivot.
It’s set to be a brilliant night.
It’s a dangerous opponent; they’re playing for it all; Tuchel understands knockouts, and could get pretty defensive in this one. Arsenal, however, are as good a team as any in world football by any conceivable metric — not to mention the eye test. There are more vulnerabilities for Arsenal to target than the reverse.
We can be overprecious about our little tactical ideas. At this level, it’ll mostly be about the players: duels, physicality, and individual moments of brilliance. As it should be.
Do you feel the approach of happiness, Anya?
Always look forward to these man, your write ups are amazing!
This was spot on. Rice was too high to stop the transition/match physicality and although Kiwior was there to do the job, we actually needed Tomi on Sane. We should have gone the Brighton away game model but instead pressed too aggressively and gave them too many opportunities to create vertical transitions. Were made to play their game and must be happy with what we got.
Consolation is that I think we’ll be a lot better in Munich because our away set up is better suited to this opponent.
Keep it coming.