An AI Clone of RuneScape Lived 24 Hours Before the Takedown: Football Has Been Running That Same Script for Three Decades

Three thousand accounts created in the first twenty-four hours. More than a thousand hours logged. Then a notice from Jagex arrived, and Elderwood Realms vanished from the internet.

I read that report on the Express Tribune on an August evening, while my second monitor was still replaying the Becamex Binh Duong versus Long An match from the 2026 season. A game built by AI had just been erased for copying another game. A few inches away on the screen, a V-League side was running the exact formation I had once written off as dead.

The two events have nothing to do with each other. But the mechanism underneath them is similar enough to make you stop.

Football has no Jagex. Football has something more efficient: an intellectual-property enforcement system that needs no court, no letter, only money, player registration rules and a league table.

What made me write this is not the game. What made me write this is that football has been doing the same thing in silence for thirty years, and almost nobody calls it by its real name.

---

Elderwood Realms: three thousand accounts, one letter, and a clean ending

Let me lay out the source facts clearly. Elderwood Realms was a browser-based massively multiplayer online role-playing game created by an independent developer named Rohan Varma. Varma asked an AI model called GPT-6 Astra inside Codex to "build RuneScape." The result was a world with locations, skilling and combat inspired directly by RuneScape, a long-running MMORPG.

The absorption rate was frightening. Three thousand accounts on day one. More than a thousand hours played inside twenty-four hours. For anyone who has shipped an independent product, those are dream numbers.

Then the Old School RuneScape community started speaking up. On the OSRS subreddit, players pointed out the overlaps and called on Jagex, the rights-holder, to act. Jagex sent a takedown request. Varma complied. He announced that Elderwood's operation would be discontinued following that notice. The website remains available only in a limited capacity, with high scores preserved.

No lawsuit. No ruling. No damages figure. Just a request, a compliance, and a product gone within days.

Now strip away the technology veneer. Three ingredients produced this collapse: a clone with a real engine, a community that spotted the duplication, and a rights-holder powerful enough to issue an order without going to court.

Football has all three. The only difference is that the process takes fifteen years instead of seventy-two hours, and we call it by gentler names: transfers, loans, academy buyouts, youth development.

---

Football is the oldest copy economy in sport

No sport copies as much as football. A pressing system succeeds in the Bundesliga and within eighteen months it appears in England, Spain, Italy, Portugal and Vietnam. An academy model produces three national-team players and ten similar projects get pitched. A broadcast format sells rights and every league wants its own version.

Copying is how football spreads knowledge. That is the good side.

But there is a line that analysts rarely state plainly: a clone is allowed to exist only as long as it does not take revenue away from the original. The moment it does, an invisible enforcement mechanism activates.

In gaming, that mechanism is Jagex's takedown notice. In football, the mechanism has no name, and because it has no name it works far better.

I wrote about this at thirty-two, freelancing for an online football site. In March 2026 I published a piece titled "Becamex Binh Duong's 3-6-1 can only paper over the season." I said plainly that coach Nguyen Thanh Son was wasting his midfield: the team averaged 612 touches per match but produced only three touches inside the opposition box. The data from five straight defeats was sitting right there in the article. Other coaches called me a troublemaker.

Then an assistant coach at Long An called me. He said he also thought the old 4-2-3-1 was dead, and that nobody dared say it in a press conference.

That was the first time I understood something that later became a working principle: When I wrote about 3-6-1, I was not picking a fight — I was describing what the whole stadium was denying.

What the whole stadium was denying, in the Elderwood case, is that a community of players voluntarily did the work of a copyright enforcement agency. They were unpaid. They simply saw a product that looked too much like the product they loved, and they reacted.

In football that mechanism exists too. But it is usually smothered by something stronger: money.

---

A tactical system as a piece of owned property

Start with the most abstract layer. Is a tactical system protected?

Legally, no. Nobody copyrights a 4-3-3. Nobody can sue a team for pressing high.

Economically, though, a tactical system is protected by a different mechanism, and that mechanism is stricter than intellectual-property law. It runs through the coaching labour market.

When Ralf Rangnick systematised pressing in Germany, his market value multiplied, and clubs paid to import the entire package of thinking. When Pep Guardiola turned positional rotation into an organising principle, a generation of coaches studied it and resold that knowledge under their own names.

The original loses nothing. If anything, the original is valued higher because clones exist.

But one group always loses this game: the club that produced the system without owning the person who thought of it. That club spent the money, ran the experiments, absorbed the losing seasons, and then when the system matured the coach was bought away, and the team had to start again with an inferior clone.

I have seen exactly that mechanism in the V-League, at smaller scale and in numbers almost nobody bothers to record.

In the 2026 season, when the pandemic forced leagues to play without crowds, I gave myself a task I considered insane at the time: collect results from fifty-six matches across the V-League and the Premier League between May and July. I wanted to know how much home advantage survives when the noise is gone.

The result made me sit still for a long time. Home win rate fell from 47.3 percent to 38.1 percent. Yellow cards for away teams rose 22 percent. I wrote a piece arguing that home advantage no longer worked the way the away-goals rule assumed, and proposed scrapping that rule. The article drew two million views in three days. Sixteen months later, UEFA abolished the away-goals rule.

I tell this not to boast. I tell it to point at something people skip past: Take the noise away and the stadium becomes a laboratory — and the home-ground myth starts to crack.

A football myth is like a long-running game. It survives on a set of assumptions never tested under clean conditions. When clean conditions appear — no crowd, no psychologically skewed refereeing, no stand pressure — the assumption cracks, and you see the real structure underneath.

For Elderwood Realms, the clean condition was an AI model that could rebuild ten years of content in weeks. For football, the clean conditions are positional data, passing data, and seasons played behind closed doors.

Both cases produce the same kind of conclusion: most of the value people assume belongs to the original actually sits in a network effect, not in the product itself.

---

Loans with obligations to buy: the takedown notice in disguise

Now the most concrete part, and the one that has drawn the most argument.

In Italian football, the loan-with-obligation-to-buy structure has become so standard that nobody treats it as a suspicious financial instrument any more. A big club takes a young player from a small club on loan, with an obligation to buy triggered after a set number of appearances. The small club gets money, but gets it late, and gets it after having paid the entire development cost.

From a cash-flow perspective, this is a perfect replica of the Jagex mechanism, except it is never called enforcement.

The financing side carries no risk. The final decision-maker is the bigger club. The developing club receives enough to keep operating but not enough to reinvest in the next generation. And in many cases the obligation never triggers; the player returns at twenty-four with a valuation set by the very club that borrowed him.

This is why I argue that loan-with-obligation deals are the single most damaging instrument for small clubs in modern football, and far more sophisticated than any direct transfer.

In Vietnam, the local version is softer but structurally identical. The Hoang Anh Gia Lai academy, built on the JMG Academy model, produced a generation any football nation would envy: Nguyen Cong Phuong, Nguyen Tuan Anh, Luong Xuan Truong, Nguyen Van Toan, Tran Minh Vuong. It remains one of the most serious youth projects ever undertaken in Southeast Asia, and it consumed years and a level of investment very few Vietnamese enterprises would ever commit.

But when that generation matured, the league structure did not let the academy capture matching value. Clubs with bigger budgets sought to sign them, or waited for contracts to expire, or applied pressure to get players below their true value. The academy kept producing. The big clubs kept consuming.

The original was taken apart piece by piece, not by a letter, but by a sequence of perfectly legal administrative decisions.

---

Former-star academies: clones without an engine

There is another kind of copying I want to name directly, knowing it will annoy people.

Over the past decade, the number of youth academies carrying the name of a former national-team player has grown fast in Vietnam. In the media, every launch is welcomed, and the stated reasons are always the same: the founder understands football, has top-level playing experience, and wants to give back.

I am not saying they are all identical. I am saying most of these models run on the logic of a clone without an engine: they copy the outward form of a European academy — pitches, kits, glossy syllabi, photos with foreign coaches — while missing the one thing that creates core value: a grassroots coaching workforce that is properly trained and paid enough to stay for the long term.

In other words, they copy the interface, not the tool.

Elderwood Realms is a very clean example of the same error inverted. There, the AI model had a real engine — it genuinely built a world that ran and pulled three thousand players in a day. What it lacked was legal ownership.

In most Vietnamese former-star academies, the situation is reversed: the brand rights exist, the technical engine does not.

What gets neglected in both cases is the same thing: a professional middle layer. In gaming, that is the intellectual-property layer. In Vietnamese football, it is the grassroots coaching layer.

Based on my experience watching matches, including youth games and fixtures in empty stadiums, the gap between a well-coached youth team and a poorly-coached one is not athletic quality. It is whether their coach was trained to teach a principle, or merely told to run a drill.

A drill can be copied in an afternoon. A principle takes ten years to implant.

Any football nation that wants to stop being a supplier of semi-finished goods must invest in grassroots coaches before investing in players. That is the only investment that cannot be copied, because it does not sit inside any syllabus you can buy.

---

Data rights: the real copyright of modern football

If any area of football is running the Jagex script exactly, it is data.

Over twenty years, football data moved from reporters' notebooks to positional collection systems measuring hundredths of a second. Data companies build models, rent subscriptions to clubs, sell feeds to bookmakers, sell charts to broadcasters. Value is created at the middle layer.

Clubs — the parties generating the raw data — capture very little of that chain. Players — the people generating data with their bodies — usually capture nothing.

In recent years some European clubs have begun demanding ownership of the data generated in their own matches. This is a genuine copyright negotiation, just held in meeting rooms rather than courtrooms.

In the V-League, this conversation is far more embryonic. Detailed match-level data is still sparse. But as positional data spreads across Southeast Asia — and it will arrive faster than people expect — the ownership question will surface. Who owns the data Vietnamese players generate in a V-League match? The club? The league organiser? The hardware provider? The player?

If the answer remains "whoever pays for the collection system," then Vietnamese football has already signed away its own ownership rights before understanding what it had to lose.

---

Community pressure: Reddit, Jagex and the Super League lesson

One detail in the Elderwood story matters more than the incident itself.

Jagex did not discover the clone. Old School RuneScape players found it, flagged the overlaps, and called on the rights-holder to act. In other words, a community performed free monitoring work for a corporation.

That model has a direct football equivalent, and that equivalent once destroyed a multi-billion-dollar plan in seventy-two hours.

In April 2026, twelve leading European clubs announced a breakaway competition. The plan had capital, media partners, and a legal framework prepared over months. What killed it was not a court ruling, but the reaction of fans and players.

On 2 May 2026, a group of supporters stormed the Old Trafford pitch during the Manchester United versus Liverpool fixture, forcing a postponement. Within days, all six English clubs withdrew.

The common thread: a gamer community on one side, a supporter community on the other, both operating as informal enforcement agencies, both succeeding in forcing owners to act or retreat.

But there is a difference I do not want to blur. In the Elderwood case, the community activated a mechanism protecting the original. In the Super League case, the community activated a mechanism protecting the ecosystem from its own owners.

Same mechanism, two directions. Which direction it runs depends on whether the community receives information before or after money has moved.

The real power of a football community lies in reaction speed, not organisation. Any project that wants to survive must either be announced after consensus exists, or executed in total silence.

---

The V-League as a quiet laboratory

Back to my second monitor.

I have spent many years watching matches almost nobody writes about. Matches in stadiums holding three thousand people. Matches in July rain in Binh Duong. Matches where touch counts were never recorded and people remember only the score.

That is where I learned the most.

In the 2026 season, when I wrote about Becamex Binh Duong's 3-6-1, what I really meant was not that the shape was wrong. What I meant was that a shape only has value when it matches execution capacity. Three central midfielders can create numerical superiority in the middle third, but if the attack produces three box touches across ninety minutes, that superiority becomes a cost centre rather than a return.

The 3-6-1 was mocked. Yet another version of the same principle — three centre-backs with high wing-backs — was being praised in Europe at the same time. The difference was not the shape. The difference was execution quality, and whether anyone bothered to record the data that would prove it.

That is what I keep returning to: the same shape, two fates, depending on whether data exists.

And that is why I treat Elderwood Realms as a football lesson rather than a technology story. The game had clear usage data: three thousand accounts, a thousand hours, in one day. It had proof of value. What it lacked was rights.

The V-League is the same. A great deal in Vietnamese football has proof of value but lacks rights, lacks legal structure, lacks the professional middle layer that turns value into protected assets.

---

Where I might be wrong

Now I have to dismantle my own piece, as I do after every mistake.

The 2026 World Cup failure taught me something I never forget: The 2026 World Cup mistake taught me that every football commentary is a chess game against yourself.

In July 2026 I was invited to commentate the France–Uruguay quarter-final on Vietnam's biggest football YouTube channel. In the fortieth minute I declared confidently that no team wins a World Cup with 45 percent possession. Uruguay lost 0-2. France were on 42 percent possession. Social media tore me apart.

I stayed silent for two weeks. I rewatched France's seven matches and found what I had missed: they needed an average of 3.6 counter-attacks to score a goal, twice the efficiency of the rest of the tournament. I wrote a three-part series using my own error as the headline. Dozens of football sites picked it up.

So where might today's piece be wrong?

First, I may be reading too much into a technology story. The line between a useful analogy and a forced one is thin. Elderwood Realms was a digital product, copyable at near-zero cost. A football club cannot be cloned that way. Stadiums, players, local identity — none of that can be rebuilt by a language model. If I merge the two, I have made a classification error.

Second, copying is not a disease. It is how football evolves. Without copying there is no Gegenpressing in Italy, no total football in Japan, no small football nation learning from a big one. The problem is not copying. The problem is the ownership terms attached to copying. If this article is read as a call against learning, I have betrayed my own argument.

Third, and this is where I must be most careful: the "underdog is always right" posture is a trap. I have spent much of my career siding with small clubs, mocked coaches, condemned tactics. But if the small club plays badly, lacks discipline, or lives on luck across several matches, painting them pretty is a betrayal of the very data language I use as a weapon.

In the Elderwood case, siding unconditionally with the independent developer would mean ignoring one detail: he asked an AI to rebuild a specific game, not to build a game inspired by it. That was a deliberate choice, not an accident.

So let me state my position clearly: what I object to is not Jagex protecting its copyright. What I object to is that football has built equivalent mechanisms that nobody is allowed to name, and the weaker parties inside those mechanisms get no chance to comply with dignity.

I am never confident in a pre-match prediction — I am only confident in my own doubt.

---

What I think happens next

I will stake one testable prediction.

Within five years, football will see its first public dispute over ownership of an AI-generated data model. Not image rights, not broadcast rights. A scouting model or an injury-prediction model trained on one club's data and sold to another without an explicit agreement.

A concrete signal to watch: when a club starts inserting a "player data ownership" clause into professional contracts, or when a league requires member clubs to hand positional data to a central entity. When either appears in the V-League, the chess game has begun.

As for Elderwood Realms, it will sit in the list of stillborn products, and in a few years nobody will remember the name. But the mechanism that killed it will not die. That mechanism is running inside every loan deal, every youth-development clause, every data negotiation happening in some office where nobody is allowed to keep minutes.

Three thousand accounts in twenty-four hours is a signal. A thousand hours played is evidence. And a takedown letter is a reminder that in both gaming and football, value is created in one place and the power of decision sits in another.

Whoever controls the middle layer between those two places writes the rules. Everyone else just plays by rules written by someone else, hoping they are not asked to take something down before they understand what they own.