Inside Riot Games' Anti-Boost Machine: The Fragile Line Between Climbing Rank and Manipulation

Inside a Diamond-ranked VALORANT lobby, most accounts perform exactly as the tier predicts. Then a few play as if the person behind the keyboard has changed. Crosshair placement, the flick at a corner, decision speed under pressure — all drift away from the average distribution of the whole rank. The feeling does not come from an opponent being better. It comes from an opponent belonging to a different tier entirely, misplaced into a matchmaking system that should not allow it. There is nothing to report. Nothing to flag. Only a vague sense that the lobby was unfair in a way impossible to prove. That feeling has a name in the publisher's documentation, and it has data behind it. Riot Games has actioned 296,416 accounts for rank-manipulation behavior across two titles, VALORANT and League of Legends, through a system called Anti-Boost. In 2026 I read Josef Martinez's xG and saw a revolution forming in Atlanta. He averaged 24 touches per match, yet his xG per shot reached 0.42, the highest in MLS. The lesson holds: numbers do not lie, only readings do. What I need from the 296,416 figure is not its size but the mechanism that produced it. Boosting is the act of a highly skilled player logging into someone else's account to play ranked matches on their behalf, earning ranked points for the account owner. Anti-Boost is Riot's automated system for detecting and penalizing that behavior alongside other forms of rank manipulation. According to documentation published by Riot Games itself, the system sorts violations into four overlapping groups: playing on another's behalf to climb, buying, selling or transferring accounts, intentional deranking, and smurf-assisted climbing. To grasp the size of the boosting market, you have to see how it operates. A Silver player who wants Diamond can pay a Challenger player to log into their account and climb for them. Behind that transaction sits a wider network: sales of pre-ranked accounts, transfers of account ownership, and deliberate deranking loops that manufacture low-ranked accounts for the next round of boosting. Every link in the chain is a money flow, and every money flow is a reason for a rulebook to exist. The technical difficulty is that boosting leaves no fixed trace. It is not injected code, nor background cheating software. It is a real person, in a real geography, holding an account that is not theirs. The system must infer intent from behavioral signals: network patterns, login timing, sudden shifts in play quality. That is indirect evidence, and indirect evidence always carries a false-positive probability. When Riot says self-operated alt accounts are normal, it concedes that the line between playing well and playing on behalf can only be drawn by inference, not direct proof. The most striking element is the safe harbor Riot draws for itself. Alt accounts created and operated by the same person count as normal activity. Anti-Boost targets the intent to manipulate rank, not the existence of alt accounts. That is a narrow, intent-based standard, very different from the blanket bans many fans imagine. The publisher chooses to strike at behavior rather than at tools — a decision with deep consequences for how the law is enforced. Based on my experience watching matches and community forums, this is where players routinely misread the problem. They think boosting is a lone cheater's act. In truth it is a supply chain with buyers, sellers and middlemen. The transfer market is where emotion gets priced, and I usually stand outside that room. This is one of the rare times I step into a similar room, except the price list is written in ranked points rather than transfer fees. Riot operates a four-tier penalty ladder, and its structure reveals more than any press release. At the first tier, once manipulation is detected, ranked points and rewards earned from cheating are cancelled, the account returns to its original rank, and a temporary suspension applies. On repeat offenses, ban duration escalates with each violation. For account trading or intentional deranking, penalties can climb to permanent bans. At the final tier, associated parties — including the booster's main account and teammates who frequently queue with them — can also be actioned. Three design decisions in this ladder shape the entire game. The system is reactive with rollback: points and rewards are cancelled after detection, meaning a lag exists between manipulation and correction. During that lag, honest players still get matched with an account being boosted. Alongside that, penalties escalate by severity, with permanent bans reserved for the most commercially driven violations. Above all, the system extends liability to third parties, turning an individual act into a multi-person event. Riot's decision to pool two titles into one disclosure also deserves a close read. The 296,416 figure spans both a tactical shooter and a multiplayer online battle arena. The two ecosystems have very different boosting economics: rank-inflation pressure, regional demand, scarcity of high ranks. Pooling hides those differences and makes any trend analysis impossible. A cumulative total is not a trend line. In transfer-market analysis, I always demand a time-based comparison before calling anything rising. In my work as a transfer-market administrator, I am used to valuing a midfielder through behavioral indicators. The same logic applies here. A rank climbed on the ladder is a certificate of amateur talent. When that certificate can be forged with money, it loses quantitative value. A scout looking at a high rank can no longer separate the person who climbed from the person who bought the climb. The result is a rise in talent-evaluation cost, and some genuine prospects get skipped because the system can no longer distinguish the two profiles. What this design genuinely does well is turn Anti-Boost into a risk-price-raising mechanism. When a boost buyer faces the possibility of a permanent ban, and a seller faces losing their main account, the expected cost on both sides rises. That is textbook economics: do not erase demand, make demand cost more than it is willing to pay. Compared with how European football federations handle match-fixing, the logic is surprisingly similar. They too cannot erase the motive, so they raise detection risk until the motive is no longer worth pursuing. But there is a blind spot the penalty ladder does not state. Every enforcement figure in the disclosure is self-reported by Riot, verified by no independent auditor. A total published by the publisher itself is a brand statement before it is a technical fact. That does not mean the data is wrong. It means the data needs a second source before it becomes the basis for any conclusion about the system's effectiveness. The counterintuitive angle sits in the joint-liability clause. Two friends who play together every Friday night, who happen to queue often with a booster, can be swept into the enforcement wave. They broke no rule. They cannot control who gets matched into their lobby. A clause that reaches broadly without a tolerance threshold or a concrete appeal mechanism is an operational ethics gap, not merely a technical detail. This is where two subjects I have followed my whole career intersect: VAR controversy in football and algorithm controversy in esports. Both circle the same question — who defines what counts as clear. The intent-based standard also creates a transparency problem. A bright-line ban, such as banning alt accounts outright, can be explained in a single sentence and needs no inference. A ban based on inferred intent requires the penalized party to trust a process they cannot see. When outcomes differ across similar cases, trust collapses far faster than when a bright line is crossed. I have watched enough seasons to know a system perceived as arbitrary loses credibility faster than one perceived as severe. Detection lag is another cost that never appears on the stat sheet. A boosting account operating for three weeks before enforcement has already shifted the ranks of dozens of other players. Their points are not fully restored, and their competitive memory of that season is not repaired. A rollback system can fix data; it cannot fix experience. Finally comes the evolution asymmetry. Riot states it will expand Anti-Boost and add match-level detection of boosting signs. Having to announce continuous improvement shows the current method is incomplete. In any race between detection and evasion, the evader holds the speed advantage, since they need to find one hole while the detector must plug every hole. Data is where I take shelter, but also where I learn to distrust every claim, even when the claim comes from the publisher itself. A clean ladder serves more than the climbing player. It is also an input to amateur scouting pipelines, where a high rank is a raw talent signal. When that signal is noisy, an entire talent-discovery funnel drifts with it, and the cost never appears in any press release. The question I leave for the next tracking cycle: will Riot publish a time-based comparison sufficient to turn a total into a trend, or keep it in the form of a proclamation. The line between climbing and manipulation will always be blurry. What can change is the transparency of how the publisher draws it.