Drinking at English Football Matches: A 140-Year Ban Is Wobbling, and Nobody Has Solved the Equation Yet

I once sat in Block B of a lower-league English stadium on a December afternoon, four degrees Celsius, and watched a man in his sixties carefully wrap his beer can in a black plastic bag — not to drink it, but to pass through the turnstile without being stopped. He performed the gesture with the fluency of a ritual repeated hundreds of times. I asked him why he didn't just finish it before going in. He shrugged, said the law was the law, and that he wasn't going to miss a match over a single can. Then he told me about his father, who had stood on the terraces in the 1960s, when beer was not yet treated as a threat to public order. That was the first time I realised that the question of "drinking at football" is not about beer at all. It is about a historical cut that England made for itself and has preserved for more than a century without anyone having the nerve to touch it.

That is not an emotional observation. It is the starting point of a model I had to rebuild, because when I read the latest analysis of the debate between the Football Supporters' Association and the UK Football Policing Unit, I realised that almost the entire argument is taking place on a plane of emotion and institutional power, while behavioural data, club-level financial data, and cross-sport comparative data are all missing. And when data is missing, people default to old fear.

The conclusion I draw before going into detail: this is not a football story. It is the story of a law written in a context that has vanished, defended by a generation of police who have never been asked to prove its premises again.

Context: A historical cut England made for itself

Let us start with the driest fact, because every debate about "supporter culture" should start with the driest fact.

In England and Wales, supporters of professional football are prohibited from consuming alcohol in areas where the pitch is visible. This is one of the oldest sport-specific laws in the world, and what is remarkable is that it applies to no other sport in the same country. Spectators at rugby league and rugby union, spectators at concerts in the very same stadiums, in the very same seats, through the very same concourses, at the very same beverage counters — they drink beer entirely legally.

This is not something I am inferring. It is the point FSA chairman Tom Greatrex emphasised when speaking on BBC Radio 4's Today programme. He posed it plainly: same stadium, same capacity, same event, but if it is rugby, beer is allowed; if it is football, beer is banned. This asymmetry in the law is the main fulcrum of the entire campaign.

And this is where data becomes more interesting than emotion. Because if you ask ten English supporters why this law exists, about seven will say "1980s hooligan violence." But if you check the chronology, you discover that the alcohol ban predates that era by a long margin. The hooligan explosion and the 2026 Hillsborough disaster generated a new, tighter layer of legislation, but the legal foundation banning drinking in pitch view existed from the nineteenth century, in the context of order concerns around late-Victorian football matches as crowds grew and grounds began charging admission.

In other words: this law was not written for the world it governs. It was written for a world most of the people defending it today have never lived in. That is the first important point I want the reader to carry.

The data analysis I am using as my foundation makes this clear: the origin of the debate is a BBC Radio 4 interview, with Greatrex representing the FSA, and the UK prime minister described as sympathetic to reviewing the law. On the other side, the UK Football Policing Unit, led by Mark Roberts, who has held the national football policing lead role since 2026, opposes change on public-order grounds.

What you need to hold in mind: these are two organisations with different functions, not interchangeable in authority, and both are telling the truth as they see it. The question is not who is right. The question is who is holding the evidence.

The data shock: Three planes the debate has not touched

When I work with probability models in football, I always begin with one principle: if a debate cannot be modelled, then every conclusion drawn from it is merely an opinion dressed in authority. The debate about alcohol in English stadiums falls into exactly that gap, on three planes.

The first plane: A causal claim without data

The central police argument, according to the recorded statements, is that alcohol plays "a significant part" in violent and disorderly behaviour at football matches. This is a causal claim. And from the experience of someone who has spent decades working with data, I can say that a causal claim without micro-level behavioural data — incidents involving alcohol, rate against total incidents, specific time and location context — is simply a hypothesis spoken in the tone of a finding.

Do not misunderstand me. I am not saying the police are wrong. I am saying that in the data analysis I am working from, there is not a single concrete figure on the correlation between in-stadium alcohol consumption and disorder incidents. No percentage, no incident count, no breakdown by division, no comparison with other alcohol-permitted sporting events.

And this is the crux that reform advocates are leveraging, legitimately in logic: if the police do not supply data to quantify their claim, the burden of proof is in the wrong place. In a legal system operating on the principle of liberty unless there is reason to restrict, the party seeking to maintain a restriction must prove its necessity, not the party seeking removal must prove the safety of removal.

I write this with the caution of a man who once staked 180 million Vietnamese dong on a conclusion drawn from xG data where I had mis-read the context. I know full well that the police's causal claim may be correct in substance even while lacking the data to prove it at this moment. But that is exactly why this debate needs a controlled experiment, not a vote of faith.

The xG shock at Hang Day turned me from a watcher of football into a reader of data. And the first lesson of a data reader is: an unverifiable claim is not a disproven fact. It is only a fact that has not yet entered the spreadsheet.

The second plane: Club-level financial benefit has not been quantified

On the reform side, the FSA argues that allowing alcohol sales in stadiums could provide a crucial income stream for lower-league clubs, which depend more on matchday revenue and have less broadcast revenue.

In direction, the argument is sound. I have spent years tracking club financial models across divisions, and the basic principle holds: the further down the football pyramid, the more the revenue structure depends on matchday, and the thinner the margins. In the Premier League, matchday revenue is a small share of total revenue; in the lower divisions, it can be a load-bearing pillar. An incremental income stream that is pocket change in the Premier League is the same stream that, at a fourth-tier club, can be the difference between survival and death in a season.

But in magnitude, the data analysis I am working from contains no figures. No average revenue a new beer counter might generate. No model of the added cost of stewarding, policing, and crowd management. No analysis of whether the new stream would be cannibalised by existing spend — whether people drinking outside pitch view, for example, would shift to drinking inside, or drink more.

And this is the point the data analysis flags as an important gap: the benefit is presented as an FSA advocacy claim, not as a financial model verified by the clubs themselves. That is an asymmetry in argumentative resourcing that I need to name before anyone uses it as evidence. In my work, when an analyst presents a number without a balance sheet attached, I always check where the number came from.

Worse, there is an offsetting cost that the data analysis notes has not been raised in the debate itself: if legalisation leads to additional security and policing requirements at precisely the lower-league clubs where margins are thinnest, the added cost may exceed the added revenue. This is not something I am asserting. It is a hypothesis that deserves modelling, and it has not been modelled.

I once built a "context coefficient" for xG after discovering that Bundesliga home teams pushed forward but generated actual xG 0.45 lower per match when there were no spectators. The lesson from that experience applies directly here: a new revenue stream does not exist in a vacuum. It exists in an operating context, and the operating context has its own costs.

The third plane: No data on what supporters actually feel

The moment that made me pause longest when reading the analysis was a very small but very important detail: the headline question of the original article was "would fans welcome drinking at football matches?" but no opinion poll answers that question.

This is a structural gap, not a minor oversight. Neither the FSA nor the police cite data on what supporters actually want. The FSA believes change would be more popular the further down the leagues you go, but that is a belief expressed as a prediction, not a measured outcome.

In my analytical work, I always follow one rule: when an important variable is left blank, the rest of the model becomes more fragile than you think. Supporter sentiment here is the blank variable. And when a variable is left blank, both sides in the debate will fill it with assumptions favourable to themselves.

Belief is a noise variable; run the emotion regression before placing the bet. In this case, no one has run that regression. Both sides are operating on an unmeasured variable, and that makes every confident conclusion a statement of belief, not a statement of fact.

The contrarian angle: The safe-standing precedent is being borrowed as an unverified card

This is the part I want to spend the most time on, because it contains the weakness and the strength of both sides in the debate, and it contains a lesson in method that I believe is more important than the final political outcome.

The FSA's strongest strategic argument is the safe-standing precedent. Four years before this debate, rail-seating standing areas were introduced in English stadiums after a process of controlled piloting. The police objected at the time, on similar safety grounds. After implementation, the feared disorder and accidents did not materialise at the predicted level.

Greatrex uses this precedent as a direct strike against the police position: if you were wrong once about the safety risk of standing, then your current argument about alcohol may be similarly wrong, and the correct mechanism is a controlled pilot, not the status quo.

This is a clever rhetorical move. And like every clever rhetorical move, it has a blind spot.

The blind spot is this: one precedent is not a statistical sample.

I want to say this very carefully, because I know this is where my view may be misread as siding with maintaining the ban. I do not side with that. I side with method.

The safe-standing precedent proves one thing only: a ban thought to be irremovable was successfully removed through controlled piloting. That is a precedent about process, not a precedent about behaviour. It does not prove that every safety ban rests on excessive caution. It proves that occasionally, such a ban rests on excessive caution.

That distinction is the whole game here. In statistics, a sample with a size of one does not allow you to estimate variance. You know one case occurred; you do not know its rate in the population. You cannot say "standing worked, so beer will work" any more than you can say "one player scored from distance, so every long shot will go in."

This is why I call this argument an unverified card. It is strong politically, weak in inference, and dangerous in expectation. If the FSA wins by leaning on it, and an incident later occurs at a match with alcohol, the public reaction will reverse faster than any model can run. That is not a prediction of outcome. It is a judgment about the risk structure of the argument itself.

And this is where the data analysis assesses very accurately: the safe-standing precedent is double-edged. It validates the FSA's controlled-pilot case, but it also hands the police a template to demand extended evidence-gathering before any change, thereby delaying reform indefinitely. Once you accept that change must travel the controlled-pilot road, you hand the timing and scope of the pilot to precisely the people who oppose it.

Kazan does not take revenge; Kazan only sets the table and waits for me to get the calculation wrong. In this case, "Kazan" is any incident at a pilot match. And once that incident occurs, the spreadsheet will not ask who was right. It will simply record, and history will read it back the way it always reads: in the safe direction.

There is another aspect of the safe-standing precedent few notice: timing. Four years between implementation and debate is a long period in modern football. Long enough to gather data, long enough to assess impact, long enough to record incidents and their non-occurrence. But no figures on incident counts or safety events in standing areas are cited in the alcohol debate, according to the data analysis I am working from. If that precedent is as strong as presented, the data from it should be on the table. The absence of that data is one of the largest gaps I note.

I have written before that a broken model is the day the data monk must burn the book back to the original scripture. In this debate, I am seeing both sides reading the same old page of scripture, each interpreting it their own way, and neither willing to reopen the original book.

Behind the numbers: The people standing on either side of the cut

I want to pause here a moment to talk about people, because a data analysis without a human at the end is an unfinished analysis.

On the FSA side is Tom Greatrex, chairman. He is not an amateur campaigner. He represents an organisation with a long history of fighting over ticket prices, safe standing, club governance, and matchday experience. When he speaks about supporters, he speaks as someone who has built many real changes in English football.

On the police side is Mark Roberts, who has held the national football policing lead role since 2026. Let that number sink in: more than a decade in the same role. He is the institutional memory of English football policing. He has seen many cycles, many seasons, many crises and many changes. He is not a simple conservative. He is someone who has accumulated a specific kind of knowledge campaigners often undervalue: knowledge of incidents that happened, not incidents that might happen.

When two people in the same industry with long careers stand on opposite sides of a debate in open collision, that says something not only about them. It says something about the structure of the debate. It tells you this debate will not be settled by simple data, because if simple data existed to settle it, two people of such competence would have found it long ago.

The data analysis describes the police position as an obstacle with reputational risk on their side if reform succeeds. That is a correct observation but needs expanding: someone who has held a role for more than a decade has more to lose than reputation from a failure. They have a decision history to defend. And that is a genuine behavioural variable, not an obstacle that can be removed by logical argument.

This does not mean the police are right. It means the FSA's model needs to account for the institutional force it is contending with, not only its argumentative force.

I recall the night of 27 June 2026 in Kazan. Germany lost 0-2 to South Korea with an xG of just 0.41, and their last six shots all hit defenders. I had published that prediction before the tournament and received hundreds of jeers. English football policing is in a somewhat similar position: they are issuing a warning that may be correct, but they lack the data to prove it, and they are under pressure from a force with far more public voice. I know that feeling. And I know that in such situations, people tend to cling harder to their position, not open it up.

Where the real risk lies

When I synthesise the layers of data and argument, this is the risk map I can build, with the corresponding levels of confidence.

Public-order risk is medium-to-high, at medium confidence. The police assert a causal link between alcohol and violent behaviour. That claim is unproven in the data analysis, but its serious nature means it cannot be dismissed merely for lack of data. An unproven claim is not a false claim.

Enforcement-cost risk is medium, at low-to-medium confidence. This is the least-discussed risk and may be the most important financially. If legalisation leads to additional security requirements at lower-league clubs, the financial benefit the FSA expects may be eroded by the very cost of securing that benefit. This is an unmodelled paradox.

Policy-reversal risk is medium, at medium confidence. If any incident occurs after change, public pressure will lean toward swift reversal. This is the risk the "it worked for standing" argument does not insure against, because standing and drinking are two different behaviours with two different risk profiles.

Data-gap risk is high, at high confidence. This is what I feel most confident asserting. This debate lacks data on all three planes: behaviour, finance, and supporter sentiment. And when a public debate lacks data on all three planes, the outcome is usually decided by who has the louder voice, not who has the better evidence.

I do not predict the future; I only read ahead the way the past continues to operate. And the way the past continues to operate in debates about sporting law is: controlled pilots win, absolute claims lose, but only after both sides have spent enormous political capital on unverifiable claims.

A note on the Vietnamese context

I write this from Saigon, for a market where football has no English-style tradition of banning beer in stadiums. In Vietnam, drinking beer in and around football grounds is part of terrace culture, not a legal exception. That creates an interesting experiential distance I want to record, not to compare who is more civilised, but to make clear that sporting law is the product of specific history, not universal principle.

When a Vietnamese supporter reads about this debate, they may find it strange. But on closer look, they will see a familiar pattern: a law written in a specific context, maintained longer than that context, becoming hard to change not because it is correct, but because changing it requires someone to bear responsibility if something bad happens.

That is a universal problem. It does not belong to England. It belongs to any system where risk is socialised and responsibility is individualised.

I have spent years working with probability models in football, and what I have learned is that every system has a point where data no longer helps, and people must decide based on values. In this case, that point is the question: how much do we value supporter freedom, and how much do we value the peace of mind of enforcement agencies, when neither can be proven with numbers at this moment?

That is a question I cannot answer with a spreadsheet. And I think that is why this debate has run for more than a century without a clear ending.

What I will track next

If I were building a tracking model for this debate, these are the signals I would put in the spreadsheet, not as predictions, but as variables to be updated.

First, any formal government announcement of a consultation or a bill. This is the signal that will move the debate from rhetoric to procedure, and procedure is always where absolute claims meet difficulty.

Second, any statement from lower-league clubs or league associations on the revenue argument. This is the signal that will supply real financial data instead of an advocacy claim. If the figure appears and it is small, the financial argument weakens. If it is large, it strengthens considerably.

Third, any evidential response from the UK Football Policing Unit on incident data involving alcohol. This is the decisive signal. If the police publish data quantifying the link, their argument gains new weight. If they continue not to publish, the FSA's burden-of-proof argument continues to gain weight.

Fourth, any proposal for a pilot at a designated match. This is the signal that the debate has entered the experimental phase, where data can begin to accumulate.

Fifth, and perhaps most important, any independent supporter opinion poll. This is the blank variable, and once it is filled, the entire debate can change direction.

In my betting-analysis work, I learned that the most important moment in any debate is not the moment someone makes a strong claim. It is the moment someone produces a number. The number ends the claims and begins the models.

Until then, both sides will keep talking, and supporters will keep wrapping their beer cans in black plastic bags before stepping through the turnstile.

The crowd leaves, the model breaks, and I learn to hear the breathing of the empty stand. But the stand is not empty in this story. It is full of people, full of noise, and full of an unanswered question: whether supporters actually want what they are said to want, and if they do, what price they are willing to pay — in money, in safety, in accepting a model where one incident can erase years of progress?

Age 59 gives me perspective: every cycle is a loop with a residual. The residual of this cycle is the question of trust that no xG table can measure. And perhaps that is what should be said loudest in a debate that, so far, has been led by old fear rather than new evidence.

I will leave one question for the next round, not to answer, but to place correctly in the spreadsheet: if England pilots, and the pilot succeeds, what will Scotland, Wales, and every other league with the same class of law learn from that data — or will they simply maintain the status quo because no one has yet been forced to prove the status quo is necessary? That is a question every sporting-law system in the world, at every level, is answering at once without knowing it.