Inside Riot's Anti-Boost Machine: 296,416 Accounts and the Fragile Border of Ranked Integrity

It always happens at night. Three in the morning in Paris, the desk lamp still on, the queue tab still open. And in the team chat window another name appears with a 78 percent win rate across four hundred matches — a number even top-tier professionals rarely reach comfortably. I used to think it was an undiscovered genius. Then I rewatched the replay and saw decisions that didn't match the reflexes of someone playing at Silver. The reaction time didn't match, the habits didn't match, even the way they released the mouse after a fight didn't match. Not a genius. An account someone else was playing. Riot Games has finally put a number on it: 296,416 accounts exhibiting rank-manipulation behavior in VALORANT and League of Legends. Not a forum rumor, not the guesswork of a skeptic. This is a figure published by the publisher itself, in a report on the Anti-Boost system — the automated engine Riot uses to hunt boosters, account traders and deliberate derankers. And as always, when a number is released, the real question isn't whether it's big or small. The question is: behind it, who is being caught, who is slipping through, and how much meaning remains in the rank millions stay up late to climb. THE BATTLE AT THE DEEPEST LAYER This war is fought at the deepest layer of the esports ecosystem: the ranked ladder. Before VCT, before international tournaments, every pro started their career with a single lonely account on the ladder, and that is where their future was decided in the years that followed. Rank is only the way people retell what they haven't understood — but in esports, it is not merely a retelling. It is money. It is a scholarship. It is a trial slot at a major academy. It is an invitation to a community tournament. That is why a black market formed: high-skill players paid to log into someone else's account and climb it to a dream rank. Others buy and sell accounts like merchandise. A third group loses on purpose to drop rank, enabling boosting or easier wins. Riot calls all of this one name: rank manipulation. And it counters with Anti-Boost — an automated detection and penalty system running across both of its biggest titles. Notably, the 296,416 figure pools VALORANT and League of Legends together, with no split by title and no split by region. The publisher does not say how many accounts belonged to the tactical shooter and how many to the MOBA. It only says the combined figure has been dealt with. That is a communications choice, and I read it as a signal. When you merge two very different black-market economies into a single number, you are not presenting data. You are presenting resolve. THE FOUR-TIER PENALTY LADDER The Anti-Boost system operates on a four-tier penalty ladder, and this structure is the part most worth dissecting. Tier one: when manipulation is detected, the system cancels all ranked points and rewards derived from cheating, returns the account to its pre-intervention rank, and applies a temporary suspension. Tier two: repeat offenses lengthen the ban, escalating with each violation. Tier three: buying, selling or transferring accounts, or intentional deranking, can lead to a permanent ban. Tier four — and this is the controversial one — joint liability: not only the manipulated account is actioned, but the booster's main account and the teammates who frequently queue with them may also be swept in. In other words, Riot does not only strike the offender. It strikes the space around the offender. I spend a lot of time looking at the boundary this system draws, because it is subtler than it appears. Anti-Boost does not ban alt accounts. Riot states clearly that creating and operating your own alt account is normal behavior. The system targets the intent to manipulate rank, not the existence of an alt account. This is an intent-based standard, far narrower than a bright line like "two accounts equals a violation." As a design choice, that is correct. Millions of legitimate players own more than one account for entirely valid reasons. A blanket ban would turn that whole population into violators overnight, generating a backlash no publisher could withstand. But what is the price of an intent-based standard? Transparency. A rule based on intent is harder to apply consistently than one based on clear behavior. How does the system distinguish an alt-account player trying to climb for themselves from someone climbing for another? The answer lies in behavioral signals: movement patterns, reflex timing, decision-making habits. And this is where I want to pause a little longer. The match begins when the coaching staff submits the roster, not when the referee blows the whistle. With Anti-Boost, the match begins at the first signal the system registers. When a Silver account suddenly shows the reaction time of a Diamond player, that is a signal. When movement patterns shift entirely between play sessions, that is another. But a signal is not proof. And the gap between those two things is the gap between fairness and injustice. Riot says it is scaling enforcement and will add match-level detection of boosting signs. That sounds like a promise about the future. But it also concedes a present reality: current methods are not perfect. The system is reactive with rollback — points cancelled after detection, rather than behavior blocked before it happens. Which means a lag always exists between the moment an account is boosted up and the moment it is dragged back down. THE MOST DANGEROUS BLIND SPOT Now to the part that keeps me up. The joint-liability clause covering "teammates who frequently queue with them" is the most dangerous blind spot in the entire system. I understand the logic: boosters seldom act alone, and hitting the surrounding network eradicates the ecosystem instead of pulling up weeds one by one. But imagine an ordinary player with a friend they climb with every night. They play hundreds of matches together, share voice chat, share tactics. Then one day that friend quietly becomes a booster — or worse, was one all along without the other knowing. The ordinary player's account can be swept into the penalty. The report offers no tolerance threshold. No minimum match count, no described appeals mechanism, no exoneration process for the inadvertently caught. In a system where the publisher is simultaneously detector, judge and executioner, the absence of an independent appeals channel is not a small detail. It is the entire problem. I am not saying Riot is doing wrong. I am saying this design places the burden on an automated system without placing beside it a proportionate check. When you concentrate the power to detect, adjudicate and punish in a single entity, you do not create justice. You create a power. And look at the 296,416 figure from another angle. It is a cumulative number with no baseline for the prior period. It does not mean the figure is rising, falling or flat. It only means that, in total, 296,416 accounts were once flagged. Yet in the report it is presented as evidence that the crackdown is escalating. That is an inference, not a fact. The data gives us a total. It does not give us a trend. People call it the meta; I call it fear, digitized. But here, the fear belongs to both sides. On the legitimate player's side, fear of being swept in. On the publisher's side, fear of losing trust in the ranked ladder — the thing that feeds the entire ecosystem above it. There is another, quieter risk: the arms race between detection and evasion. Riot scales enforcement, but boosters evolve too. When easily detected channels are sealed, activity flows to harder-to-detect ones — organized deranking rings, communication over external platforms, more sophisticated account-sharing patterns. The report admits detection must keep improving. That means the current system is designed to run a race, not to win once and forever. WHAT I WANT TO SEE IN THE NEXT DISCLOSURE I look at this number not as a victory, but as a mirror held up to something every competitive ecosystem faces: the value of rank comes from trust, and trust cannot be coerced with bans. What I want to see in the next disclosure is not a bigger number. I want a rate of resolved appeals. A tolerance threshold for joint liability. A quarterly comparison point. Because if all we get are ever-growing numbers without a self-correcting mechanism, then what we are building is not a clean ladder but an obedient one. And an obedient ranked ladder has never been the place where legends are made.