Theory of Soccer: A world without tactics

Like if Jurgen Klinsmann coached all the teams

So yea, imagine then that there is football but no “football tactics.” What does this soccer world look like? This is problematic from the get-go because defining terms is not my strong suit. But maybe we can just start with.. imagine that there are no coaches, that all the teams just roll the ball out there and play. We should imagine also that the players don’t huddle up with a leader emerging from the chaos, himself setting the formation and assigning players to different roles. Imagine that no one does that. But hold in your head the idea that players have natural tendencies and effectively “positions” that they’re used to playing which they learned at some point because I said so. This isn’t entirely coherent, right? As John Muller and I discussed on Episode 7 of the Post Script podcast, the one about Brian Phillips’ Run of Play and his cycle on Football Manager, there’s really no such thing as a “no tactics.” The players that go out there and play will always take with them the things they’ve learned about the game. There’s always gonna be some kind of latent tactical vibes going on. Even still, relax your eyes until you see this. It is a useful starting point for us: a kind of base state of nature for soccer.

State of Nature

A “State of Nature” is a concept you’ll see often in certain branches of philosophy. It’s a very uh “classical one” and it’s bad, but people often frame things up (economics, law, politics) in these discreet classical ways, so I want to try this on for size, and we’ll just see where it takes us, always keeping in mind that this is a mistake. In their posts, Thomas Hobbes and John Locke and many others variously imagine a state of nature where men are basically free to govern themselves based on reason (*giggling*).

Philosophers contemplating politics and statehood often use this idea of a state of nature (rarely properly credited as a fictional thought experiment) as a starting point from which “man” finds a reason to form a state. He basically finds that everyone governing themselves is cool and kinda badass (doesn’t apply to women), but there would be less suffering if they all gave up some of these freedoms to form a government. The natural result of this philosophical framing is that since people today are often influenced (directly or subconsciously) by philosophical history, political disputes are often framed as disputes around how much natural freedom (and what kind) should an individual sacrifice to the State (an exogenous one I guess) from his starting natural point of total freedom, and what sorts of guarantees/rights must a government acquiesce back to individuals in exchange for moving away from full freedom towards governance.

This classical state of nature is a fiction because man is born into society, period, whether within the paradigm of nation states or not. And when you start from that origin of “man already in society,” your analysis lands you at a more heterodox radical spot, a more progressive, inclusive, and caring one than when you start from a classical perspective of “1. I rule. 2. I am smart. 3. If I’m expected to care about others or be governed by them, what will I get in exchange?”

Anyway, we’re not gonna talk about politics and stuff (unless that’s all we’re ever really talking about when we talk about soccer). We need this fictional state of nature for soccer (no tactics), as a launching pad conceptually towards tactics and its place in a theory of soccer (you can already feel this fraying at the seams I bet). How does soccer get from this fictional state of nature (one without coordinated tactics but instead just some sense of reason) to what we know of as the modern tactical approach. Maybe we try this on for size and then.. some time from now, we might end up at some sort of fundamental description of soccer (inclusive of tactics) that doesn’t need the fictional state of nature as an origin story.

Fictional Tactic-less Soccer State of Nature

So what does our tactical state of nature looks like?

The players in our thought experiment will go out there and just broadly know what types of things to look for (and worry about) during play despite no top-down allocation of roles and responsibilities. What they won’t do is coordinate team-wide in very “tactical” ways. Because it’s a team sport (an inherent brokenness to this analogy, or is it?) I would propose that players in our fictional tactic-less state of nature will broadly sing from the same hymn sheet in terms of covering for each other on defense, sharing the same ideas around prioritizing which spaces are most dangerous for their opponents to occupy the ball in, and share similar ideas around the need to stop the opponent from progressing the ball towards their goal. What they would not do is collectively press high or collectively maintain a compact deep defensive block. They may not even share the same ideas around what types of triggers to look for in terms of when to press and when not to. To the extent we have to squint and picture the team instead of the collection of players, it’s one that is neither proactive nor reactionary, but a secret third thing. The players are improvising. They have some kind of intuition. They’re aware of each other. It’s like an itch they can’t scratch, but they’ve gotta take care of it somehow.

In attack, you might imagine that they have an intuitive sense that they need to help each other progress the ball forward. They show for passes, they run in behind, they look to combine in uh “normal” ways. Your mileage may vary, but when I try to picture their improvisation, I don’t imagine them as being so hell bent on helping their teammate on the ball to protect and progress it that they would group up near the ball abandoning all sense for their general “positions.” Surely they would feel some anxiety about doing that, because they have a general sensation that to leave their general territories would be to break open their own team’s shape, becoming vulnerable to counter attacks when they inevitably lose the ball. These things aren’t memorialized anywhere, they are intuited. Something nags at them. Man and his reasoning. (I feel tensions already, how is it that they have enough of a sense of some kind of team shape that they feel anxiety about breaking it, unless they’re coordinating in some way? When they do “normal” defensive cover, is it zonal? is it man? When they defend, do they want to win the ball back, or is that something that will just naturally happen later?)

They probably don’t have a shared idea of how to handle goal kicks either — do they build from the back, does the keeper kick it long. Perhaps the keeper directs his team in the moment to take what the defense “gives” him while I admit I’m sort of saying the defense doesn’t know what to give him in the first place. Still, we need this fictional state of nature, in order to play around with and frame up the transition from it and towards “tactics.”

I squint at it and I kind of see a fairly back and forth game if the teams are relatively even (hold that thought). Nobody’s pressing super high, nobody’s battening down the hatches. Nobody’s kicking it long the whole time, nobody’s tiki-taka-ing from goal to goal trying to pass it in. You end up with a just kind of “blah” cloud of tactics (lower case T) - feels like an open game to me but your mileage may vary. And of course that blah cloud of tactics isn’t actually anything strictly natural, it’s full of biases I’ve constructed, full of incoherent tendencies on the part of the players that come from me deciding what is collective action and what isn’t…. pretending some things come from “pure reason” while other things must come exogenously from the coach or who knows, maybe generatively from some other source of wisdom.

Now, consider what this soccer state of nature looks like if 1) the teams are _not_ relatively even. What does it look like if one team is good and the other is bad and if everyone’s just operating by feeling out there (again not on a solely individual level, but also not on a top-down coordinated level either). Everyone’s operating on instincts, skill, and athleticism, and natural ability to team with one another and one team has a lot more of “the right stuff” than the other team. It should result in one team creating many more chances than the other team (you can picture this on a stat sheet as xG or xT or shots ratio, or you can take a deep breath and just kinda daydream about what it looks like). But you get one team outplaying the other in concrete ways (whether or not the dice rolls go their way or not). That’s the basic state of nature I picture between unequal soccer teams without coordinated action.

Typing this, I suddenly see a clear image. It’s the opening act of Stanley Kubrick’s “2001: A Space Odyssey”: The Dawn of Man. There’s a roving pack of apes that either through a rather abstract show of strength and courage (they yell really loudly and jump up and down), invades a territory with a watering hole that another roving pack of apes (our proto-protagonists) relinquishes in retreat. None of these apes are necessarily coordinating from a top down perspective, they’re feeling it out but definitely divided into two different packs here. The retreating apes are left in a miserable situation, sort of hiding in a cave at night, looking depressed, perhaps cold… probably thirsty, but not for long. Next, Kubrick has the retreating pack of apes wake up one morning to find the iconic “Monolith,” and something important changes for these apes.

They become inspired and begin to jump up and down, shouting, then one at a time (improvising you might say) they all work up the courage to touch the thing. Some time after that, one of the apes is just kind of sitting there looking at the skeleton of a previously expired mammal, now a pile of bones lying on the ground, and he suddenly gets !an idea! *orchestra readying themselves for a fanfare*. He picks out one of the long bones lying on the ground and holds it in his hand. He begins to smash the other bones lying on the ground as Kubrick needle drops some Strauss.

Jurgen Klopp: “We watched this very boring video, 500 times, of Sacchi doing defensive drills, using sticks and without the ball…. We used to think that if the other players are better, you have to lose. After that we learned anything is possible; you can beat better teams by using tactics.” (Michael Cox’s Zonal Marking)

In 2001, what that one ape learned (perhaps we’re to imagine this is his first true idea) is that even if the other apes are stronger/braver, you don’t have to lose. You can beat the other apes using tools. Offscreen between scenes, the ape provisions long bones as weapons to his compatriots and then in the next scene they together take back the previously surrendered territory, using these bones as weapons to disperse the other pack of apes through both fright and harm.

In what is widely considered the best edit in cinema history, one of the inspired apes in a celebratory impulse takes the leg bone he’d been using for violence and hurls it above his head. The bone rotates freely through the air, the camera (close in) tracking it upward until it slows to where gravity starts to overtake the toss. Finally, still rotating in the air the bone begins to fall back down to earth but right at that moment Kubrick cuts seamlessly to a shot of a starship floating weightless in space as it begins a docking maneuver with a rotating space station. For the audience it’s dizzying and exhilerating. The subtext is unmistakable. While this cut jumps us forward in the story thousands of years chronologically, the more important function is to establish as the central theme of the film: the evolution of man and civilization by way of inspiration and invention.

Apes found the exogenous variable of the monolith, something in their brain was transformed, and they gained the capacity to imagine a bone as a tool/weapon. This moment (The Dawn of Man) kickstarts the history of human invention and thought which (as a film editor is want to do) can be fast-forwarded to the point where humans are now flying starships in space as the Blue Danube rings out triumphantly.

Similarly, when you read that Jurgen Klopp quote, you can almost hear the gain on a guitar amplifier start to build while Klopp stares into a camera and says “after that we learned anything is possible” and then as the guitar feedback crescendos “you can beat better teams by using tactics” *POWER CHORDS.* If this is a documentary, we’d see a black title screen with “Heavy Metal Football” in white text as Japandroids’ “The Boys are Leavin’ Town” blasts across all speaker channels.

Klopp took over FC Mainz while they were floundering at the bottom of the second division in Germany. He put into place his famed gegenpressing tactics and the team won 6 out of 7 and narrowly avoided the drop to the third division. Within two years, Mainz were promoted to the first division of the Bundesliga. They finished midtable their first year in the Bundesliga despite having the smallest budget in the league. Two seasons later they were relegated. The following year they didn’t achieve promotion, and Klopp resigned.

*bone falls to the ground*

The history of soccer isn’t as straightforward as the rapid evolution of man, aided in part by the mysterious monolith and in part by the Deus Ex editing of the director.

This is because the history of soccer exists within the history of [everything else]. Accordingly, the history of soccer exists within the *ahem* material history of global capital. Basically, unless you have already, through various forms of consensus… coordinated to form a soccer league (or a soccer federation) that enforces principles of relative competitive parity amongst its clubs, you end up with leagues that are instead set up to automatically and pro-cyclically re-enforce this “market” tendency of rich clubs to get richer and poor clubs to get poorer, and not because of some grand/global conspiracy — the opposite: because the more successful a club is, the more money it makes, and (so long as a club is vaguely bounded by its business results) the more money it makes, the better players it can hire to play for it, with the reverse being true as well. I mean at least that this is the “natural” world a soccer club can be thought of as existing in given the material conditions brought on by centuries of unnatural intervention and obfuscation by these very ideas of “natural mechanics” peddled by *gestures broadly* at the expense of *gestures broadly* brought on by not so much a grand conspiracy as a series of unfortunate improvisations (and yeah, some conspiring)

Stay focused for fuck’s sake.

So while, Klopp is almost definitely right in a narrow sense — that through tactics you can beat better teams as he showed finishing midtable with the lowest payroll in Bundesliga — there’s a real practical limit to this phenomenon. And if you are able to beat a team with more resources than you, you’re just as likely to do so via a lucky roll of the dice than by outflanking them strategically. And you might just as well outflank a team strategically while also rolling snake eyes and lose. Further, if they’re significantly better than you, you might outflank them and still lose straight up. Anyway, all of this adds up over time such that the teams with the highest payrolls tend to do better than everyone else. At the margins there are some interesting questions, but the model as a whole seems to be quite clear. The teams with the better players win.

But what about those margins? What if we wanted to explore those more? If we re-do the thought experiment above, and instead of picturing teams that are unequal in resources (the so-called “state of nature”), we somehow put into place *sunglasses on* a grand conspiracy to ensure relative parity across the teams, well what now? *amplifier gain growing* How might that look? What might that platonic ideal of soccer look like?

*electric feedback screeching* Well, some of us live in this Edenic paradise week in and week out, as we follow, watch, and analyze a little thing called:

*power chords*

Major League Soccer is amazing. I love it, and I hate it. (Why do I do this? Nesciŏ, sed fierī sentiō et excrucior). In Major League Soccer, they try to make the playing field relatively balanced while not forcing it to be perfectly even. There’s a salary cap that’s centrally funded by the league that the 30 player roster has to fit within, but for 2-3 players on that roster, the individual teams can go hog wild — say sign the greatest player to ever put on boots (at costs mostly born by the owners of said individual teams). And in the locker rooms, the boots of these designated players are often sat right next to those of college graduates who make $100K a year. What ends up happening is that some teams do in fact spend more than others, but on the whole the gap is a lot smaller than we see in “big time” European leagues. If the highest payroll in the Premier League is 10x the lowest team, and the highest payroll in La Liga is 24x the lowest team, and the highest payroll in Germany is 27x the lowest team, then Major League Soccer’s highest team payroll is closer to 3x that of the lowest team. Quite often in Europe the worst player on one team might be better than the best player on the opposing team. This doesn’t happen in MLS. In any match.

That’s a big deal. Compared to other major soccer leagues, the parity in MLS is nearly perfect. Up there at the beginning, we were talking about how at the margins, surely tactics can make an impact, and if that’s right, then Major League Soccer lives and breathes those margins. It’s 100% those margins with nearly all the talent inequity excluded by identity. So in this universe of tight margins, do tactics decide the results of matches and seasons of matches in Major League Soccer? Is MLS a manager’s manager’s league? Hell no. Not in any way that we can observe at least.

There are a bunch of other factors that impact the results of matches. I talk about these things a lot. But the most obvious ones are:

1. Home Advantage

2. Randomness (inclusive of injuries)

Home Advantage in MLS

In the major European leagues, home teams tend to score on average about 0.45 points per game more than they do away (for any poor non-soccer folks in here, 3 points for a win, 1 point for a draw). For the absolute best and worst teams, this difference is smaller, (say +0.2 ppg for Arsenal & City, -0.2 ppg for Sheffield, Burnley). Basically, if you’re very good or bad, then it doesn’t matter where you play, the impact of home advantage drowns in the whirlpool of other factors (such as how good your players are). Conversely for teams that are similar in quality, you would expect the home advantage to be higher: if it’s hard to otherwise separate two teams, then a one-sided advantage like home field should do the trick as the number of matches increases.

In Major League Soccer, home teams tend to score on average about 0.7 points per game more than they do away. And the “best” teams in MLS do similarly! (In 2023, Cincinnati and St Louis who won the East and the West respectively were closer to +0.8 at home). Of course, the “best” MLS teams are very similar in quality to the “worst” MLS teams (on an absolute scale). So we see a more evenly distributed home advantage (and a much larger one overall!). MLS Stadiums are further apart than stadiums in England, but the crowds are hardly more rabid. If Premier League teams were suddenly similarly salary capped, my guess is that home advantage would balloon to as high as +1 PPG due to the peerless home matchday atmospheres (which is the primary transfer mechanism of home advantage - crowds influencing calls).

With a home advantage so large, if tactics are driving the outcomes of MLS matches at the margins, then it has to get in line behind other factors like this one.

Unpredictability in MLS

While all sports with balls include randomness and variance, a good starting point for predicting a team’s future performance is usually that team’s past performance (or performance to date). One of the tasks of analytics then is to find what sort of metrics of past performance or performance to date do a better job of predicting future performance. One of the nice findings of the early analytics stuff was that since there were more shots in games than goals, adding goal probability to the shots data gave you a more powerful sample with which to predict future goal scoring. In fact, it’s powerful enough to predict future goal scoring better than past goal scoring does. In Europe, halfway through a season, performance to date metrics using xG predict future goal difference with an r^2 of just under 0.5. Soccer is still penty random, players have free will, and there’s of course room for variance. The past is just the past after all, but that 0.5 is kinda nice given the nature of the sport. Especially when you contrast it to how well xG does in MLS:

charts courtesy of Eliot McKinley who wrote about this for American Soccer Analysis. This makes me LOL every time.

Using shot based data (like xG) still gets you more predictive power in MLS than using past goal scoring data, but the total predictive power of anything in MLS is just very weak (r^2 of 0.2 instead of 0.5). Whatever the margins are between the teams, they’re very hard to replicate from the past to the future in MLS. Since they’re running around trying to control a spherical object with their feet, any measures of performance to date just don’t correlate well with those same measures of performances in the future, including goal difference. It takes a huge talent advantage to consistently overcome this level of messiness. In a salary capped league, who’s to say that anything that’s already happened in an MLS season is very predictive of what will happen, when what has already happened has been so contingent on randomness?

What about tactics?

So if we do our best to remove the talent disparities away from soccer (shrink this gap using MLS as a proxy), it’s not so simple that what remains to separate teams’ performances is a variation in team tactics. It is not the case that MLS is a tactics sandbox where league titles are determined perennially by X’s and O’s, the sort of thing you see in minor college football markets that are always birthing the next big tactical thing.

In theory tactics are somewhere in there in soccer - yes different teams use different tactical approaches (in MLS you’ll see teams press and possessing, you’ll see man and zonal defending, you’ll see bunker and counter, you’ll see back 4s and back 3s, diamonds, and false 9s, and flat 442s), but I just struggle to find any evidence that it’s impacting the results of current matches in a significant way amidst all of the other more material factors we know about today. And it’s not clear that further refining the parity somewhere like MLS will somehow land us in a place where tactical differences rise to the surface and determine the outcomes.

You could look at coaches in MLS and try to identify where the consistent tactical success is I guess? Wilfried Nancy whose possession-heavy game model has smitten many in the soccer media landscape lifted the Cup in 2023 with the Columbus Crew, but so did Caleb Porter in 2020. Porter was then fired by the Crew after his team missed the playoffs in two consecutive seasons and while he’s back this year at New England, he’s hardly getting the kind of respect you’d expect for a coach who’s tactical tweaks had helped two different franchises lift silverware. Ronny Deila lifted the cup in 2021 with New York City FC, but hasn’t lasted more than one season in two stops since (relieved of his duties early at Standard Liege and Club Brugge). For every success story, there’s normally a healthy dose of “front office cleverness” mixed with “good old fashioned luck” to taste. Last year Minnesota United had the sort of expected goal difference that should’ve landed them top 4 in their conference (MLS has an Eastern and Western conference) and hosting a playoff match, instead they… finished 11th in their conference and stayed home. In 2021, Bob Bradley’s LAFC led the West (and joint led the entire league) in xGD but finished 9th in the conference, missing the playoffs, and the coach and club parted ways. It’s a weird league! The highest win percentage all-time in MLS amongst coaches that played 8 or more games is Frank de Boer, who was run out of Atlanta in the middle of his second season, and has famously failed quickly at every club he’s coached at not named Ajax.

For the sport as a whole, we know that conceptually, talent level is the main driver. And we see it poke through and impact results even in the face of significant home advantage and random variance. We also see that as parity increases, the effects of talent discrepancies become more muted. Results become less predictable. And we also know that as parity increases, steady structural factors such as home advantage can be observed at higher levels. But tactics? Like how the hell would it even show up? John Muller recently tried to find something similar in his article “How can you tell if a football manager is actually good at their job?” He’s not isolating for “tactics” per se here — after all a manager’s job is much larger than “team tactician” — but it shows how tricky parsing one thing from another in a cloud of balls bouncing can be. If tactics matter, everyone seems to be doing a good enough job such that it’s not as if the tactical contributions of certain managers are observable in their results (and others not).

The Paradox of Skill

This phenomenon where because the teams in MLS are more equal, luck determines the outcomes more than it otherwise would is akin to what Michael Mauboussin calls the “Paradox of Skill:”

The key is this idea called the paradox of skill. As people become better at an activity, the difference between the best and the average and the best and the worst becomes much narrower. As people become more skillful, luck becomes more important.

He’s mostly imagining a dynamic where there’s some (again) “state of nature” (some imagined free market) where there are discrepancies of skill that are easily observed and normally distributed between participants. As this population of participants improves via investment in training, learning from failures, access to shared knowledge etc, the difference between the best and average and worst narrows, and so what was once more of a skill-based meritocracy (clear observed differences were cemented in outcomes) becomes more of a roll of the dice (competitive edges are worn away). Mauboussin is often (or ultimately) talking about investing when he references this paradox of skill, and we can sort of short-circuit or short-cut it by flipping from European soccer where skill is gobbled up by capital creating real discrepancies in talent to Major League soccer where relative parity is enforced (rather than some natural improvement in skill that brings everyone closer together).

But what about coaching and tactics? The paradox of skill might be more directly applied to coaching. Because in the absolute sense, coaching has a huge impact on the outcomes of matches, right? If a coach tells one team to stand on their own goal line, and this is the tactics that they all execute, it may not matter how much more talented those players are than their opponents. They will lose. But in a realistic sense, no coach would put those tactics into place. At the professional level of football, the skill of designing and helping your players to execute tactics may very from manager to manager, but in relative terms, it might be nearly equal, or equal enough such that these discrepancies in skill cannot routinely overcome the other factors (talent, luck, home advantage).

One reason for this narrowness in coaching skill is the observed process whereby over time tactical knowledge is shared, observed, tested, trained etc.

But there is also just this natural “put option” on manager/tactical impact. There is a “floor” to the impact of differences in tactical skill between teams which is simply that if there was no coach, the players would sort of figure out a way of playing together that was adequate, such that the differences in player quality would still overcome things — but only if the best players were out on the field, and the coach picks the players. So the floor and ceiling of player quality difference (and roster selection) is much lower and higher, respectively than the floor around specific tactical choices, bounded on the one end by the absence of a manager and at the top end by some sort of Borg-like mind meld of all reasonable tactical ideas across all the players in perfect rhythm.

Carlo Ancelotti: There are two types of managers, those that do nothing, and those that do a lot of damage. The game belongs to the players.

It’s basically impossible for a manager at the professional level to ruin a team with poor (in an absolute sense) tactics. It is very possible for a manager at the professional level to refuse to allow the players to play tactics that fit their strengths, and it is very possible for a manager at the professional level to not pick the best team. In terms of choosing tactics, most managers at the professional level can adequately assess their player roster and the types of roles the players should play and therefore the types of tactics necessary to empower those players. If they can’t, they should go. But it’s hard to find evidence that some managers possess tactics that are themselves differentiators (they get better outcomes) from other managers’ tactics. To the extent that these do arise, they are swiftly assimilated by other managers such that the tactics themselves return to that state of not making a dent on the results compared to other macro-factors.

I think that’s what I want to write about in the next post. Some brief histories of tactical changes. Maybe we can assess when these matter and how, and for how long? And where do they even come from anyway? Is there even such a thing as a beginning state of nature? What is football tactics?

Ps: We lost someone special last week way, Om Arvind of Tactical Rant and many other places.

Please consider donating to a good cause in his memory. https://gofund.me/6d477cd7