296,416 Accounts and One Prevention System: Riot Games Rewrites the Rulebook of VALORANT and League of Legends Ranked Play

In the final week of the most recent ranked season, I sat down with my tracking board and a single number published by Riot Games: 296,416 accounts exhibiting rank manipulation behavior across VALORANT and League of Legends, counted from late last year to the moment of disclosure. I read that number three times in one evening. Not because it was large — though it truly was — but because the structure behind it tells a story that the Vietnamese and Korean esports communities rarely discuss. When a publisher publishes a total figure, they are not bragging about an achievement. They are signing a memorandum with their own players: this is what we define as a violation, this is how we detect it, this is the penalty, and this is the limit of punishment. The question is not whether Riot is tightening the screws. The question is whether the system they have built can withstand its own logic — and I believe there is at least one fracture point, located in a clause almost nobody noticed.

The context I want to reconstruct before diving into the analysis is not the context of a match. It is the context of a power structure. Riot Games operates two of the largest ranked ladder systems in the global gaming industry: VALORANT, a 5v5 tactical shooter, and League of Legends, the MOBA that has now lived for more than a decade. Both of these systems share a characteristic I always emphasize when talking to sports-analysis students in Seoul: the ladder is not just a place to sort skill levels. It is the talent supply source for the entire professional tier above it. A young Vietnamese player hoping to be noticed by an academy in the LCK or VCT Pacific almost always begins by climbing to Challenger or Grandmaster on the Korean server. When rank is manipulated, the tier above is poisoned at the root.

That is why I read the material on Anti-Boost not as a technology news item, but as a governance document. Over the past two weeks, I spent eleven evenings translating and cross-referencing information from Riot's side, while comparing it with what I observed in the Vietnamese player community on the Southeast Asian server and a few groups of Vietnamese players climbing the Korean ladder. What I saw did not lie in the number 296,416. It lay in how that number was broken down — or rather, in how Riot chose not to break it down in the way analysts need.

Riot Games operates the Anti-Boost system as a punishment machine based on intent, not on the mere existence of alternate accounts — and this is the core difference that most players misunderstand. I have seen many comments on Vietnamese forums saying "Riot bans alt accounts." That is technically wrong. Riot draws a clear line: an alt account created and operated by the player themselves is normal activity, permitted. What is targeted is conduct with the intent to manipulate rank — climbing by playing on someone else's account, buying or selling accounts, or intentionally losing to drop rank.

This demarcation matters far more than it appears. When a publisher moves from a rule of "ban X" to a rule of "ban conduct X for purpose Y," they shift from a bright-line standard to an interpretive one. And every interpretive standard opens two doors at once: the door protecting legitimate players from wrongful punishment, and the door for inconsistency in enforcement. Both doors are open in the current design of Anti-Boost, and I will point out the specific fracture point later.

Riot's penalty ladder is designed in four tiers, and I want to present it as an escalating structure rather than a list. Tier one: when the system detects manipulation, rank points and rewards derived from cheating are cancelled, the account is returned to its original rank, and the owner receives a temporary suspension. Tier two: on repeat offense, the ban duration increases — an escalating penalty mechanism. Tier three: account buying/selling or intentional deranking can lead to a permanent ban. Tier four — and this is the tier I want to spend the most time on — associated parties may also be actioned: the booster's main account, and even teammates who frequently queue alongside them.

I want to pause on tier four. In six years of observing the esports industry from Seoul, I have seen many punishment mechanisms built, amended, and dismantled. And the lesson I have drawn repeatedly is this: every joint-liability punishment system pays for its scope with accuracy — the more you expand coverage, the more you increase the probability of catching innocent people. The clause "teammates who frequently queue together" is logically sound as a preventive measure — because there are pairs who coordinate to boost together — but it lacks two things every joint-liability punishment system needs to function fairly: a specific pairing threshold, and a clear appeal path.

I tried to calculate the probability of this scenario in my head. If a normal player, entirely unaware of the business dealings of their queuing partner, plays 50 matches with a friend over one season, and that friend is in fact boosting for another account, then the probability the innocent player gets swept into the enforcement list — under current logic — sits at a level I estimate around 30 to 40 percent, depending on the frequency of shared queuing the system records. This figure does not come from Riot-published data — Riot does not publish it. It comes from the structure of the clause. When you define a behavior as "frequent," you need a quantitative threshold. When you don't provide one, you hand the decision to an algorithm. And the algorithm, at present, cannot distinguish between the friend who queues 40 matches out of intimacy and the friend who queues 40 matches out of ladder convenience.

Data tells a story the media does not have the patience to hear. In this case, the data Riot provides is a single figure with no prior-period baseline. I searched through official announcements for a figure from any previous period — any period at all — to construct a trend line. There was none. The number 296,416 is a cumulative total. It tells you the scale of the problem at a point in time, but not whether the problem is contracting or expanding. In sports-industry data analysis, I always apply one principle: one data point is not a trend; a trend requires at least three points across a time series.

This is where I want to turn to the contrarian angle. The esports community, in both Vietnam and Korea, reads Riot's announcements in an almost uniformly single direction: "Riot is tightening." Headlines all revolve around the word "tighten." But when you look at the enforcement structure rather than the number, you see a different story. What Riot is actually doing is not tightening — it is widening coverage. These two things sound similar but differ in essence. Tightening is raising the penalty for the same conduct. Widening coverage is adding new conduct to the violation list, and adding new subjects to the enforcement list. Riot is doing more of the second than the first. And widening coverage without increasing detection accuracy is a strategy with a ceiling.

You see, the Anti-Boost system operates on a reactive-with-rollback model. Points and rewards are cancelled after detection, not before manipulation occurs. This means there is always a lag between the moment the violation occurs and the moment it is stopped. During that lag, the manipulated account can still climb, still participate in ranked matches, still influence the match outcomes of other players. And in a ranked system with resonance characteristics like VALORANT or League of Legends — where one skill-disparate player can decide the outcome of an entire match — this lag has non-trivial real-world value.

I once wrote about a similar topic in the context of football: how millimeter offside lines kill attacking instinct. The logic here bears a notable resemblance. When a control system is designed to make the game more precise, it inadvertently changes how the game is played. The referee becomes the editor of the match. In the case of Anti-Boost, the algorithm becomes the arbiter — and that arbiter faces a problem I believe cannot be perfectly solved with behavioral data: distinguishing intent to manipulate from coincidence in play patterns.

An empty stadium is not because the audience is absent, but because belief left before them. I use this line often when speaking about public-relations crises in sports, and it applies here too. When a normal player is swept into an enforcement list because they queued with a friend, what they lose is not only rank points. They lose faith in the consistency of the system. And faith in consistency — not in severity — is what sustains a healthy ranked system.

I want now to move into the section I call risk diagnosis before solutions, following the method I have used since my analysis of FC Seoul's doll scandal in 2026. When I analyzed that case, I split the crisis into three layers: operations, communications, and fan trust. I predicted brand recovery would require at least 14 months. In the Anti-Boost case, I also see three risk layers, but ordered differently.

The first layer is the economic risk of the gray market. Anti-Boost, by heavily penalizing account buyers and sellers with permanent bans, directly attacks the supply side of the account market. In theory, this raises the expected cost of participating in the market — if detection probability multiplied by loss magnitude rises, the expected value of boosting activity falls. This is the basic economic logic of any deterrent. But I want to emphasize something I learned from my tracking model: gray markets do not disappear when suppressed — they migrate to harder-to-detect channels. In the boosting case, this means activity moves from in-game communication channels to external platforms, from public groups to private ones, from personal accounts to coordinated networks. This makes Riot's detection problem harder, not easier.

The second layer is the risk of community perception. When you publish a large number, you create an expectation of a large problem being handled. But that expectation has a shelf life. If in the next disclosure the number does not rise significantly, the community will interpret it in two directions: either the problem has been solved, or the system is hiding data. Both directions carry costs. And if in the next disclosure the number spikes, the community will interpret it in a third direction: the system is failing to prevent. This is a classic communications trap — any subsequent number carries risk.

The third layer is institutional risk. Riot operates both the detection system and the adjudication system, with no independent appeals body described in the documentation. Governance authority is fully concentrated in the publisher's hands. This is not wrong legally — Riot owns the platform and has the right to set its own rules. But structurally, it creates a situation analysts call "authority-concentration risk": when one entity writes the law, enforces the law, and adjudicates violations of the law, the system's capacity for self-correction depends entirely on that entity's goodwill.

I realize my analysis thus far may read as criticism. I want to be clear: it is not. In the esports industry, I have observed enough publishers to know that those willing to spend money on a prevention system like Anti-Boost are very few. Most publishers choose to ignore the problem, or handle it minimally to placate media pressure. Riot, by building a prevention system with its own name, with a tiered penalty structure, with a joint-liability model and an expansion roadmap, is investing in a commodity I call "legitimacy of the ranked system." This is a long-term investment, and it can be measured by a single indicator: the rate at which players return to ranked each season.

But investing in the right direction does not equal an optimal design. And this is the point I want to spend the rest of the article clarifying: what should be fixed, and what should be tracked.

First, on the intent-based standard. I understand why Riot chose this approach. A blanket ban on all alt accounts would destroy the experience of a large number of legitimate players — those who create a second account to play with lower-ranked friends, or to experiment with different play styles without affecting their main rank. A blanket ban is the easy but wrong solution. The intent-based standard is the right but hard solution. And when you choose the hard solution, you must accept that you need to invest more in transparency mechanisms. Transparency here is not just publishing the number. It is publishing how you distinguish intent, how you handle edge cases, and how players can protect themselves from being swept into an enforcement list.

Second, on the liability threshold. A clause of "teammates who frequently queue together" without a quantitative threshold is a clause that cannot be externally audited. Players do not know where they stand relative to that threshold. There is no way for a duo to self-assess their risk. In system design, this is a form of "black box" — users bear the consequences of a rule they cannot observe. I believe Riot will soon have to publish a specific threshold, or a self-check mechanism so players know where they stand. The probability of this happening within the next 12 months, in my assessment, is about 70 percent.

Third, on the match-level detection problem. Riot has mentioned expanding detection of boosting signs at the match level, rather than only at the account level. This is a technical step forward but also a potential step back in accuracy. Account-level detection usually relies on relatively objective signals — IP addresses, devices, login patterns. Match-level detection relies on behavioral patterns — and behavioral patterns always contain noise. A player in a training phase with a new role may show behavioral patterns resembling a booster. A player undergoing psychological difficulties, performing worse than usual over a streak, may trigger signs of intentional deranking. I am not saying these signals are unhelpful — they are helpful. I am saying that moving from account signals to behavioral signals will increase the probability of false positives, and Riot needs to prepare for that with an operational appeals mechanism.

Speaking of appeals, this is the biggest gap in the current documentation. There is no description of an independent appeals mechanism for liability-related enforcement cases. In any automated punishment system — from traffic fines to bank account locks — the existence of an operational appeals mechanism is the decisive factor for maintaining long-term legitimacy. It is not a concession to violators. It is a self-correction mechanism of the system.

Now, I want to move to what I think is the most valuable part of this analysis: the effect of Anti-Boost on the esports ecosystem by layer. In the transmission model I usually use — upstream is the publisher, midstream is the ranked system, downstream is player experience, and peripheral is derivative markets — Anti-Boost produces different effects at each layer.

At the upstream layer, Anti-Boost is a trust-maintenance investment. It protects the legitimacy of the two ranked systems, and thereby protects the daily-active player base — the foundation of the entire esports talent funnel. This is a positive but slow effect. You will not see its signs in a quarter, but in two or three years.

At the midstream layer, the effect is more complex. Heavily penalizing account buyers/sellers and boosters reduces the supply of the account market, but may simultaneously push service prices higher — because supply falls while demand has not yet fallen correspondingly. This is an effect I call the "price paradox of deterrence": when you make a violating activity harder, you may inadvertently make it more valuable to those willing to accept higher risk. I do not have data to confirm this effect in Riot's specific case, but it is a pattern I have observed in other gray markets, and it warrants tracking.

At the downstream layer, meaning player experience, the effect is clearest but unevenly distributed. Compliant players receive indirect benefit from a cleaner ladder. But players swept into a liability case — even once — suffer direct losses in points, time, and trust. The ratio between these two groups, by my estimate, tilts heavily toward the first. But the esports media industry has a notable characteristic: it amplifies individual cases far more strongly than aggregate trends. One wrongful enforcement case that spreads will create an impression of a systemic problem, regardless of its actual frequency.

At the peripheral layer, Anti-Boost exerts downward pressure on gray transaction flows, including activities related to derivative betting — where account rank is sometimes used as an evaluation signal. I want to be clear: the documentation does not indicate any direct link between Anti-Boost and betting. This is my inference about an indirect transmission chain. And like any indirect inference, it should be treated with corresponding uncertainty.

There is one more dimension I want to address, and it is the least discussed: the scouting value of a clean ladder. For many years, one of the most important talent supply sources for professional esports has been very-high-rank solo-queue accounts. Academies and scouts use rank as a primary filter before detailed VOD evaluation. When the ladder is manipulated, this filter loses value. An account at Challenger no longer guarantees that the person behind it truly reaches Challenger level. And for a young Vietnamese player trying to make a mark on the Korean server, this means the signal they send becomes harder for scouts to read. Anti-Boost, if operating effectively, will restore the signal value of rank. This is a benefit the documentation does not mention, but in my view it is one of the most valuable long-term benefits.

I return to my tracking board after three weeks. I have a list of signals to monitor. First, Riot's next data disclosure — if there is a prior-period figure for comparison, I can construct the trend line I currently cannot. Second, the emergence of a high-spread wrongful-enforcement case — this will be the first real-world test of the legitimacy of the intent-based standard. Third, a clarifying statement on the liability threshold — if this appears, the risk of catching innocent players drops significantly. Fourth, a disclosure by another publisher of a similar anti-boost system — this will provide comparative context for the scale Riot is publishing. Fifth, signs of the account market migrating to harder-to-detect channels.

Success on the pitch is recorded in goals, but its cost is recorded in other numbers. In the esports case, the cost of a well-functioning ranked system is recorded in a different number: the number of players who never have to wonder whether their opponent is truly at their skill level. That is a number nobody publishes, because it measures the absence of a problem. But it is the true measure of success in governance.

Leaving the pool is not giving up; it is movement when you know the old water has limits. I left the youth swim team due to a shoulder injury at 13, and switched to recording 17 matches of the U15 Suwon Samsung Bluewings. The lesson I carried from that period is: you cannot improve a system you do not measure. Riot is at the early stage of serious measurement. The number 296,416 is a starting point, not an ending point. And the question I carry into the next evening, when I reopen my tracking board, is not "how far has Riot tightened" — but "how does this system measure itself." A publisher willing to publish data on punished players should also be willing to publish, in the same place, data on wrongfully punished and restored players. When those two numbers appear side by side, we will know the system has matured. Until then, all we have is a total figure, a penalty ladder, and a gap in exactly the position where the self-correction mechanism should sit.