Riot Games' Anti-Boost: When the Publisher Is Cop, Prosecutor and Judge

A Match at 2:47 a.m.

At 2:47 a.m. Seoul time, I stopped on the 34th minute of a ranked match no one but me had any reason to rewatch.

The account on screen belonged to a Gold II player, 340 games that season, a 48 percent win rate. This match sat in a Diamond III lobby. Across the first twelve rounds he landed 41 kills and died seven times. I had to slow playback to 0.25 speed to clearly see his crosshair sitting half a beat ahead of every corner — the kind of reflex no Gold II sustains across three consecutive rounds, let alone twelve.

What made me stop was not the skill. It was the pattern. That account only played above its level when queued with one specific friend. The other eighteen games that week, he played exactly to his Gold standard — some of them worse than his own average.

About a week later, Riot Games published a cumulative figure: 296,416 accounts actioned for rank manipulation across VALORANT and League of Legends, covering a window described only as from late last year to the time of publication.

I sat still for a while after reading that line. Not because I was stunned. Because I realized Anti-Boost had been quietly running long enough to count into the hundreds of thousands, while I — someone who has tracked esports for nearly eighteen years, who has been wrong and right often enough to know the value of evidence — had never read a single document fully describing how it works.

Every generation needs one shock to believe the impossible is possible. But the shock here did not come from a play on a big stage. It came from an administrative name: Anti-Boost.

Context: A System of Judgment With No Courtroom

Boosting is when a highly skilled player logs into another person's account and plays ranked matches on their behalf, so that account climbs to a tier its owner could not reach alone. The paying party buys status. The carrying party sells skill during spare hours. In between sits an account, a password, and a ladder that believes everything is happening fairly.

The technical reality behind that matters for readers who do not play games: a ranked ladder is not entertainment software. It is a database of human capability. Every match records a behavioral sample, and the matchmaking system uses that sample to decide who meets whom next. When an account is played by someone else, the data is contaminated. The real player in the following match faces opponents stronger than they should. The innocent player on the other side loses a game with no way to understand why.

Anti-Boost is Riot Games' automated detection and enforcement system for that kind of data contamination.

According to Riot's published information, the penalty structure runs across four tiers. Tier one: on detection, all ranked points and rewards obtained through cheating are cancelled, the account is returned to its original rank before interference, and a temporary suspension applies. Tier two: repeat offenses escalate ban duration. Tier three: buying, selling or transferring accounts, or intentional deranking — deliberately losing to lower one's own rank in order to enable boosting or easier matches — can result in a permanent ban. Tier four, and this is the part worth the most attention: the booster's main account, along with teammates who frequently queue with them, may also be actioned.

Riot also carves out an explicit safe harbor. Self-created, self-operated alt accounts are normal activity. Anti-Boost targets intent to manipulate rank, not the existence of alt accounts.

Two facts about the source must be recorded up front, because they shape everything that follows. First, Riot published no specific date for this policy update; the only time frame given is the relative window from late last year to the time of publication. Second, the 296,416 figure pools both VALORANT and League of Legends, with no per-title and no per-region breakdown.

Those two details sound like administrative trivia. They are not. They are the entire story.

Core: Five Layers of a Machine With No Appeal

A Penalty Ladder Designed Like a Price List

The first thing an analyst notices reading Anti-Boost's penalty table is that it is not written as a moral code. It is written as a risk price list.

Look at the tiering. Temporary suspension covers isolated manipulation. Permanent bans are reserved for the two categories with the clearest commercial motive: account trading and intentional deranking. The booster who rents an account to climb for someone else receives a lighter sentence than the account trader. The person who deranks their own account for money receives the heaviest.

That order is not arbitrary. Riot is pricing by damage to data infrastructure, not by moral depravity. A single boosted account distorts a small pool of matches over a limited window. A freely traded account market distorts the entire traceability of player identity across the system — and that is damage no rank rollback can undo.

From a governance standpoint, this is a coherent design choice. From a player standpoint, it signals that the system prioritizes protecting the publisher's data asset ahead of protecting any individual player's experience.

The Intent Standard and Its Trap

Riot's decision to exclude self-operated alt accounts from enforcement is one I rate highly on technical grounds. Banning all alt accounts outright would destroy a legitimate behavior: players who want to practice a new agent without tanking their main rank, or who want to play with friends at a lower skill level.

But the price of that choice is that Anti-Boost must adjudicate on intent. And intent is the hardest thing to prove transparently in any rule system.

Compare it to match-fixing law in football. There, the violation usually attaches to a near-binary signal: a bet either exists or it does not; a payment either landed in an account or it did not. The line is clear enough that a sports tribunal can rule on it. Anti-Boost has no such luxury. An alt account shared by multiple people in one household looks identical in signal terms to an account being boosted: same IP address, same device, wildly fluctuating skill by match.

Riot states clearly that not every alt-account use is actioned. But it publishes no tolerance threshold, no appeal mechanism, no false-positive rate. An intent-based standard, run automatically, with no independent adjudicating body behind it, creates exactly the gray zone that every judicial system in the world is taught to narrow.

Joint Liability: The Sharpest Edge in the Document

This is where I want to spend the most space.

The clause stating that teammates who frequently queue with a booster may also be actioned is one of the harshest governance measures I have seen at the publisher level.

As anti-cheat logic, it has a basis. In many boosting models, the booster does not operate alone. They queue with a fixed group to reduce risk, to mask behavioral patterns, to share accounts and split revenue. Punishing only the carried account means the system is forever chasing a shadow.

As governance, it opens a door I am not certain Riot has fully measured.

Imagine an ordinary player, entirely unaware, who happens to queue with a booster three or four times in a week because matchmaking places them in the same time slot. No money changes hands. No agreement exists. No evidence shows they knew they were being carried. But their pairing pattern matches the pattern the system uses to identify accomplices.

In traditional sports, joint liability exists but always travels with process. A football federation seeking to convict a player as an accomplice to match-fixing must establish knowledge, must hold a hearing, must grant the accused the right of rebuttal before a panel whose members sit outside the organizing body. Here, none of that is described. The publisher detects, the publisher concludes, the publisher enforces.

I do not believe Riot is abusing this. I believe the current structure makes abuse hard to detect if it occurs. And in governance, hard to detect is often more dangerous than occurring at all.

Retroactive, Not Preventive

Anti-Boost, as described, is a post-violation system rather than a pre-violation prevention system. Points are cancelled after detection. Rank is rolled back after climbing. Rewards are clawed back after collection.

This model carries a structural lag between the moment of manipulation and the moment of remediation. During that lag, hundreds of matches have been played on contaminated data. Real players lost. Real players dropped points. And when the system reverses points for the boosted account, it cannot reverse the loss for everyone who met that account.

This is not a design flaw unique to Riot. It is the physical limit of every pattern-detection anti-cheat system. But it is a limit worth stating, because the way it is communicated often creates the impression that the machine runs ahead of violators, when in reality it runs behind.

The Asymmetry Between Detection and Evasion

Within Riot's own published information sits a detail I reread several times: the stated plan to expand Anti-Boost's scope and add match-level detection of boosting signals.

A publisher does not announce plans to improve detection if current detection is good enough. Saying you are building additional match-level detection capability is an indirect admission that current capability lacks resolution.

This is the nature of an arms race. Pattern detection relies on behavioral data. Boosters adjust behavior to produce different data. They slow down in high-risk matches. They fragment their game counts. They move to off-platform communication. They route through intermediary accounts. Each new detection measure pushes violating behavior into a harder-to-recognize shape, and that shape demands another measure.

Recidivism rate is the only metric that shows who leads this race. Riot's document contains one indirect signal of it: the existence of escalating penalties for repeat offenses. If recidivism were near zero, escalation rules would be unnecessary. Nobody builds a staircase for a behavior that never repeats.

The Problem With a Pooled Total

296,416 accounts, pooling VALORANT and League of Legends.

These two titles run entirely different boosting economies. VALORANT is a tactical shooter where individual skill expresses itself directly and visibly in every duel; a strong player can single-handedly decide matches across many tiers. League of Legends is a multiplayer contest where an exceptional player is still bounded by four teammates; boosting power here is more diffuse and more role-dependent.

Rank-inflation pressure differs between the two systems. Market demand for boosting differs. How players price their rank differs.

Pooling two titles into one total means we know how many people the system caught, but not what kind of pressure it is fighting, where, or at what intensity. As communication, a big total lands hard. As analysis, it is a data point that cannot be broken down into action.

Based on my experience tracking matches and regional ladders, boosting demand is extremely uneven in distribution. It depends on whether rank converts into money, into career opportunity, or into social status within a community. A region where high rank is a ticket into academy pipelines runs a totally different boosting economy from a region where rank is just a badge.

Riot does not publish a regional breakdown. That leaves the most important question — where the damage is worst — unanswered.

Self-Reported Data and the Trust Problem

When I worked at a sports data analytics firm in Seoul, I learned something that became a professional principle: a data point is only as strong as its verifiability.

The 296,416 figure is self-reported by Riot. There is no independent auditor. No third-party cross-check. No period split for comparison.

That does not mean the figure is wrong. It means the figure is unverified. And in esports analysis, the gap between those two states is the gap between a report and a press release.

In 2026, I built a predictive model combining sensor data from K League footballers with win probabilities for League of Legends matches. My model missed the 2026 LCK Summer final, where Gen.G Esports lost 0-3 to Damwon Kia. The cause was not the algorithm. The cause was that I ignored a variable no data could capture: the psychological pressure of competing in total silence.

When the stands are empty, you hear your own breathing clearly — that is where every tactic begins.

I wrote a 5,000-word self-rebuttal after that failure. The lesson still applies to Anti-Boost: a system measuring human behavior that refuses to acknowledge the limits of its own measurement will always project more precision than it has.

Contrarian Angle: What 296,416 Does Not Say

I want to open this section with a confession.

When I read headlines about Riot's crackdown, my first reaction was belief. I believed because I wanted to. I had watched ladders long enough to see accounts climb in ways skill could not explain, and I had waited years for a system-level response.

But belief does not die on the day the match ends; it dies when we stop asking questions.

And the first question must be: what data do we actually have?

We have a cumulative total. We do not have the prior period for comparison. In statistics, a single figure does not make a trend. To claim Riot is tightening harder, we need the number from six months ago, a year ago, two years ago. If this total accumulated over years at a steady enforcement rate, it tells a very different story than accelerating enforcement.

Without that data, the claim of tightening is the writer's interpretation presented as an event.

The second question concerns power structure. Riot designs the rules, builds the detection tooling, issues rulings, and executes sentences. No independent appeals body is described. No panel with members outside the publisher. No published tolerance threshold for the teammate-liability clause.

In traditional sports, such concentration of authority has existed and has produced governance scandals lasting decades. Craig Lord, the leading investigative journalist in world swimming, called the conscience of his sport by The Times, spent much of his career showing that when an organization both governs and adjudicates, its decisions tend to serve its own survival first.

Riot is not a sports federation. It is a company that owns the entire ecosystem: the game, the servers, the ranked system, the professional circuit. That gives it the ability to act faster than any federation. It also gives it the ability to define fairness without anyone's agreement.

The third question, and the one I consider most important: if rank is a product, what is boosting?

A ladder has value only when players believe in it. When that belief is strong enough, rank becomes a social currency. And social currency always generates a black market. Boosting is not a disease attacking the ladder from outside. It is a parasite born from the very value the ladder was designed to create.

That does not mean boosting should go unchallenged. It means this fight has no endpoint, only a degree of control. A publisher can raise the cost of violation, but cannot eliminate the motive, because the motive is fed by the system they are protecting.

And here is the point I want Riot to address more clearly than any other: in a system where rank determines scouting opportunity, academy invitations, roster offers, a contaminated ladder does not just ruin the recreational experience of ordinary players. It distorts the talent-discovery pipeline of an entire industry. A scout looking at high rank to find prospects may be looking at an account played by someone else.

Nothing in Riot's published information addresses that consequence. But by ecosystem logic, it is the heaviest consequence an unclean ladder produces.

In football and esports, the one thing that cannot be staged is the moment belief collapses.

Takeaway: Memory Frozen at the Data Layer

I still keep that 2:47 a.m. clip on my drive.

Not because it is beautiful evidence. It is an unremarkable match at an unremarkable tier, between players no one will remember. I keep it because it reminds me that every large governance system is built from hundreds of thousands of small moments like it: an account playing impossibly well for twelve rounds, a teammate who does not know they are being carried, a player on the other side losing a game without understanding why.

Anti-Boost will not tell those stories. A total cannot tell any story. But it is a starting point for knowing where we stand.

Viewers can leave, but the stories we tell will stay at the arena.

The first shock is never a mistake; it is an invitation to rewrite the story. And the story 296,416 accounts invite us to rewrite is a story about power: who gets to define fairness, who gets to judge, and who gets the right to appeal when the system is wrong.

Riot has answered the first two questions very clearly. The third remains blank, and in an ecosystem where rank is currency, leaving it blank any longer is a choice — not an oversight.

If tomorrow Riot publishes a tolerance threshold for teammate liability, and an independent appeal mechanism for wrongful bans, their system will be stronger than all their anti-cheat patches combined. That is what I am waiting for. Not a bigger total, but a total that can be challenged.