Anti-Boost: Riot Games Sanctions 296,416 Rank-Manipulating Accounts and the Trust Equation

"In an empty stadium, I heard my own voice more clearly than ever."

I wrote that line in the summer of 2026, when every competition on the planet froze at once and I had to build a podcast series about stands with no one in them. Four years later, reading Riot Games' statement on its Anti-Boost system in a café in Seoul, I noticed something similar unfolding: some matches are staged with nobody watching, and precisely because nobody is watching, they are the easiest matches in the world to bend.

Football has referees, VAR, dozens of cameras, and a crowd large enough that anyone can become a witness. The ranked ladders of VALORANT and League of Legends have only an algorithm. No stands. No broadcast tape. No one standing up to shout when something looks off. An account can climb from Silver to Diamond without its owner ever touching a keyboard. The winner is not in the chair. The loser never knows who just beat them.

Riot Games calls this behaviour boosting. In its latest Anti-Boost update, the company announced it had actioned 296,416 accounts engaging in rank manipulation across both VALORANT and League of Legends. That is a striking data point, and also a strange one: no comparison period, no split by title, no split by region.

In other words, we know the scale of the crackdown. We do not know the scale of the problem.

---

A ladder is a market, and markets have sellers

To understand why this matters, you have to accept that a ranked ladder is not merely a scoreboard. It is a market, and every market runs on supply and demand.

In South Korea, where I live and work, a Diamond or Master account carries real social value. It is a screening criterion at many esports academies. It is a filter on job applications at some companies. It is something people put in their self-introductions. When an asset has social value, it soon acquires monetary value. And when it acquires monetary value, it breeds a grey economy nobody governs.

Riot defines rank manipulation across four behaviour classes. Boosting is a high-skill player logging into someone else's account and playing on their behalf. Buying, selling, or transferring accounts is the second class. Intentional deranking — deliberately losing to lower your own rank — is the third. The last is using a secondary account to assist illegitimate climbing.

There is an important detail here that often gets skimmed. Riot does not ban alt accounts. That is a deliberate line. A player creating a second account and playing on it themselves is normal activity, even part of ranked culture. What is targeted is the intent to manipulate rank, not the existence of an alt.

This approach differs substantially from some other publishers. Many take the simpler route: restrict alt accounts, or require phone verification to cap the number of accounts per person. Riot took the harder road — distinguishing behaviour rather than distinguishing tools. That is a defensible choice in principle, but it pushes the system into territory where everything depends on reading intent.

And reading intent is the hardest job in any enforcement system.

In 2026, when I was 19 and first allowed into the press area at a K League match between FC Seoul and Jeonbuk Hyundai, I learned something I have applied ever since to reading publisher statements: the most important thing is rarely in what people say out loud. That day every camera pointed at the ball, while I noted the Jeonbuk coach's signals on the touchline. My read proved correct and Jeonbuk won 2-1. My male colleagues laughed when I first offered the take. After they reviewed the tape, they stopped.

"Where I was once doubted is now where I find my answers."

Reading Riot's statement works the same way. The interesting part is not 296,416. It is that Riot chose to define violations by intent, and that it inserted a joint-liability clause.

---

Boosting is not new, but it has become a profession

Before getting into the mechanics, the issue deserves its proper historical frame. Boosting has existed almost since League of Legends introduced ranked play. In the early years it was small-scale and relationship-based: a skilled friend played a few games on someone's account in exchange for a meal or a thank-you.

By the 2020s it had become an organised industry. There are sellers, brokers, escrow intermediaries, refund policies if an account gets banned, and after-sales warranties once the climb is complete. Some operations run shift schedules, divide territories, and maintain rosters of repeat clients. In other words, it professionalised faster than any enforcement system could keep up with.

When an illicit activity professionalises, three things follow. Operating costs fall, so prices drop and reach more customers. Internal competition pushes groups toward more sophisticated evasion. And a class of people emerges who make a living entirely from the activity — and those people will not leave the market merely because penalties got harsher.

That is why I do not read 296,416 as a sign of victory. I read it as a sign of scale. You do not action nearly three hundred thousand accounts if the problem is a handful of small groups.

---

The four-tier penalty ladder: punishment designed as an economics problem

Anti-Boost runs on a four-tier penalty ladder, and notably each tier targets a different class of violation rather than simply escalating severity.

The first tier is rollback. When manipulation is detected, all ranked points and rewards obtained through cheating are cancelled, the account is returned to its original rank, and a temporary suspension is applied. In essence, this punishment does not penalise the past. It erases the results of the past.

The second tier is escalation for repeat offences. Ban duration increases with the number of violations. This mechanism exists for a very practical reason: without it, the first penalty becomes a fixed cost that people are willing to pay in exchange for rank. Any enforcement system lacking escalation will be neutralised by the offender's own resources.

The third tier is permanent bans, reserved for the most clearly commercial behaviours: account trading and intentional deranking.

The fourth tier is joint liability. The manipulated account is actioned. The booster's main account may also be actioned. And players who frequently queue with them fall within scope.

These four tiers tell a fairly clear story about how Riot ranks severity. Boosting is treated as a violation requiring repair and deterrence. Account trading is treated as a violation requiring elimination. Logically, that ordering holds: account traders run a business, while boosters are merely selling labour.

But there is another structural feature worth stressing. Anti-Boost is a reactive system, not a preventive one. It detects after the behaviour has occurred, then rolls it back. That means there is always a lag between the manipulation and the remedy.

During that lag, clean players matched with a boosted account still eat the loss. They still lose points. They still lose time. And nothing compensates them. Rollback restores the state of the manipulated account; it does not restore the state of those who played alongside it.

Very few write-ups mention this, yet it is the real cost of reactive anti-cheat.

---

The joint-liability clause: where a technical system meets social limits

The most contested element in the entire Anti-Boost description is the clause actioning players who frequently queue with a boosted account.

In theory it is sound. If you climb alongside a booster for many games, odds are you know what is happening. Actioning companions also blocks a workaround: punish only the manipulated account, and people shift to a two-person model, one booster playing and one legitimate partner providing cover.

In practice, there are cases where a player randomly queues with a boosted account for a handful of games, knows nothing, and gets swept into a penalty. The statement specifies no queue threshold. No appeal mechanism. No method for distinguishing those who knew from those who did not.

I do not think this is a minor detail omitted from a summary. I think it is a design gap.

When a system cannot prove intent, it is forced to infer intent. And when it infers, it will be wrong at some rate. That rate may be tiny — but for a system applied to hundreds of thousands of accounts, a tiny rate is still a very large number of people.

That leads to a question every sports regulator has faced: when you punish the innocent, you damage the very reason you exist.

---

The trouble with 296,416

Back to that figure.

It is self-reported by Riot, unaudited by any independent party, with no comparison period. It tells you total scale, not trend. The source article contains a claim that Riot is tightening enforcement. That is the writer's inference, not data.

A single cumulative figure cannot establish an upward trend. To say this year is tighter than last, you need two points of reference. Here there is one.

There is a further problem: Riot pools VALORANT and League of Legends into one number. For communications purposes, that produces a more impressive total. For analytical purposes, it destroys comparability.

VALORANT is a 5v5 tactical shooter where individual skill can hard-carry many situations. League of Legends is a MOBA where a strong player still depends heavily on four teammates. Boosting incentives, service pricing, rank-inflation pressure, and account-market structure differ between the two titles.

Pooled together, we learn the combined scale of a crackdown across two entirely different fronts. We do not learn which front is harder.

---

The arms race between detection and evasion

Riot says it is expanding Anti-Boost and adding match-level detection of boosting signs. That is a sensible direction, and also an admission.

It admits the current method relies on behavioural signals, is not accurate enough, and needs upgrading. It admits account-level detection is insufficient, and that the system needs to read the signals inside the match itself: movement tempo, decision patterns, how an account responds in specific situations.

This is the nature of an arms race, and I have watched it repeatedly in traditional sport. Every time a system recognises a new behavioural signature, the other side adapts. Communication moves off-platform. Actions get fragmented to resist recognition. Organised deranking rings appear, running on schedules, shifts, and contracts.

There is no endpoint in this race. Only speed.

And that is why I do not underestimate the seriousness of the problem. An enforcement system works only as long as it is faster than its target. When it falls behind, it does not merely lose effect — it breeds a sense of injustice among compliant players who watch violators climb unbothered.

---

The grey economy: hitting the supply side, but what about demand

The boosting economy runs on familiar logic. Demand comes from people who want a rank they cannot reach themselves: to qualify for an amateur tournament, to polish a profile, to satisfy ego, or simply to avoid being flamed by teammates next game. Supply comes from high-skill players who want extra income.

Here lies the detail the source leaves open, and the one I find most worrying. The phrase "high-skill players" may include professional or semi-professional players. If so, disciplinary exposure at the professional tier is real — and it has not been spelled out in any Anti-Boost document.

Heavy penalties on account buyers and sellers strike the supply side of the grey market. That is a reasonable choice, since supply is easier to identify than demand. But it does not address the origin of the problem.

If demand does not fall, harsher penalties simply raise the price of the service. Boosters price the added risk in. Buyers pay more. The market does not vanish; it becomes more expensive and migrates to harder-to-detect channels.

I am not arguing Riot should stop punishing. I am arguing that a supply-side-only strategy will always be one step behind the problem it wants to solve.

---

A lesson from traditional sport: doping and match-fixing

The parallel is impossible to unsee.

When European football federations tightened doping control, they quickly discovered something: testing catches only those still using, and those tend to be amateurs or the under-resourced. The most sophisticated systems always stay a step ahead, with usage cycles calculated to fall outside the detection window.

Match-fixing behaves similarly. Major rings are not uncovered because one match looks strange. They are uncovered because betting networks look strange, or because an insider speaks. The match itself usually looks clean.

Boosting in esports sits at the intersection of both models. It has the technical character of doping, because the player is using an ability that is not theirs. And it has the organisational character of match-fixing, because it runs through networks with division of labour.

So when I read about Anti-Boost, I do not expect a complete technical solution. I expect a system designed to fail continuously at an acceptable rate, and to be continuously repaired. That is how every sports-integrity system actually operates.

---

The overlooked consequence: the ladder is scouting infrastructure

There is an indirect consequence the source does not mention, and to me it is the most important in the long run.

The ranked ladder is an input into talent identification. Esports academies, teams, and scouts all look at high rank to find players. In South Korea, where the academy system is dense and highly competitive, a Master account is sometimes the first ticket to a tryout.

If high rank is diluted by boosted accounts, that signal loses value. A scout cannot separate the genuinely good from those who paid to look good. Screening costs rise. And as screening costs rise, so does the number of missed talents.

This is why I consider the fight against boosting — however small it sounds inside a server room — one of the most important infrastructure investments a publisher can make. It is not glamorous. It does not build tournaments. It merely keeps the measuring stick measurable.

But a measuring stick is only worth something if people believe in it. And belief cannot be enforced by penalties.

---

A centralised governance model: efficient and risky

Riot owns both titles, operates both ranked systems, detects, adjudicates, and publishes on its own. There is no independent adjudication body. No third-party audit of the numbers.

This is a fully centralised governance model. It has an obvious advantage in speed: a decision can travel from the data-analytics room to enforcement without intermediaries. In a race decided by speed, that is a genuine edge.

It also concentrates all error risk in one actor, with no mechanism to correct that actor when it errs.

Compare traditional sport: major federations have appeals mechanisms, sports courts, independent review. Those mechanisms do not exist because federations are weak. They exist because federations understand that unchecked power loses legitimacy over time.

Riot describes no appeals mechanism in its Anti-Boost communications. It may simply be omitted from the summary. It may also be a real gap.

---

Where I could be wrong

I want to use this section to name the places my argument could collapse.

The joint-liability clause may be more precise than I assume. Riot may apply a very strict technical threshold — for instance, actioning only accounts that queue with a booster across a significant share of games over an extended period — and that threshold may be high enough to eliminate nearly all accidental cases. If so, my concern is overstated.

The self-reported figure may also be more trustworthy than I credit. Riot has strong incentives to publish honestly, because if it is caught misstating and that surfaces, the reputational damage far exceeds any benefit from inflating an anti-cheat metric.

And there is one more possibility I must concede: the boosting problem may not be as large as it is framed. A figure of 296,416 sounds dramatic, but set against tens of millions of active accounts, the ratio could be a very small fraction. If it sits at a few per thousand, this is a well-managed issue, not a crisis.

I raise these possibilities not to retract the argument. I raise them because analysis that cannot name where it might be wrong is just propaganda with numbers attached.

One thing I hold firm on. An enforcement system built on inferred intent, applied to hundreds of thousands of accounts, with no public appeals mechanism and no independent audit data, cannot yet be called complete. It may be effective. It cannot yet be called trustworthy.

---

What I carry with me

"The widest stadium is not where the crowd is largest, but where people are willing to listen."

I thought about that line as I finished the statement. Traditional sport has stands, spectators, and social pressure forcing institutions toward transparency. Ranked esports has none of that. Only a publisher and its players, with an algorithm between them that nobody is allowed to inspect.

Riot is doing more than most publishers. Publishing the numbers at all, even unaudited, is a step many never take. But the next step is not a higher count of actioned accounts. It is showing players how the system works, and giving them a path to object when it is wrong.

Because a ladder is only credible when people believe it is fair. And trust, unlike rank, cannot be bought, cannot be boosted, and cannot be reinstated by a rollback.