Trang chủEsports296,416 Accounts and the Fatal Flaw in Riot's Anti-Boosting Rules

296,416 Accounts and the Fatal Flaw in Riot's Anti-Boosting Rules

**Câu trả lời cốt lõi**: Riot Games vận hành hệ thống Anti-Boost tự động nhằm xử lý hành vi cày thuê và thao túng thứ hạng trên VALORANT và League of Legends, với thang hình phạt bốn tầng, tiêu chuẩn dựa trên ý định và trách nhiệm liên đới mở rộng tới đồng đội thường xuyên xếp hàng cùng. **Dữ kiện chính**: - 296.416 tài khoản bị đánh dấu vì thao túng thứ hạng trên VALORANT và League of Legends, theo Riot Games. - Bốn nhóm vi phạm: cày thuê, mua bán hoặc chuyển nhượng tài khoản, cố ý tụt hạng, leo hạng bằng tài khoản phụ. - Tài khoản phụ tự tạo và tự vận hành không bị xử lý; Anti-Boost nhắm vào ý định thao túng thứ hạng. - Mua bán tài khoản hoặc cố ý tụt hạng có thể dẫn tới cấm vĩnh viễn; tái phạm làm tăng thời hạn cấm. - Tài khoản chính của kẻ cày thuê và đồng đội thường xuyên xếp hàng cùng cũng có thể bị xử lý. **Nguồn**: Riot Games (thông báo chính thức về hệ thống thực thi Anti-Boost); tài liệu gốc không nêu ngày công bố cụ thể. **Hỏi đáp liên quan**: - Hỏi: Hệ thống Anti-Boost có cấm tài khoản phụ không? Đáp: Không, Riot Games chỉ xử lý khi phát hiện ý định thao túng thứ hạng, không xử lý việc sở hữu tài khoản phụ. - Hỏi: Người chơi không vi phạm có thể bị xử lý liên đới không? Đáp: Có thể, nếu được hệ thống xác định là thường xuyên xếp hàng cùng một tài khoản cày thuê, và Riot Games chưa công bố ngưỡng cụ thể. - Hỏi: Con số 296.416 có được kiểm toán độc lập không? Đáp: Không, đây là số liệu do Riot Games tự công bố và chưa có bên thứ ba xác minh.

In the enforcement data that Riot Games published, there is a number so round that it invites instant belief: 296,416 accounts flagged for rank manipulation across VALORANT and League of Legends. But the thing that made me sit back and reread each line slowly was not that number. It was a subordinate clause buried among the penalty sections, almost never quoted: players who "frequently queue with" a booster may also be actioned.

I read that sentence four times. Four times, because I could not believe my eyes — not because it was shocking, but because it was familiar. Nineteen years in this industry, from esports competitor to tournament organiser in Vietnam, to sitting in Guangzhou writing for the Chinese market, and I have seen this exact pattern repeat in every sports rulebook ever written: people write rules to catch violators and accidentally touch people who violated nothing. Riot just did that, at a scale of nearly three hundred thousand accounts.

Context: when the ranked ladder becomes a marketplace with buyers and sellers

To understand why that clause matters, you have to understand what Anti-Boost is and what space it operates in.

Anti-Boost is Riot Games' automated enforcement system targeting boosting — a high-skill player logging into someone else's account to climb ranks on the owner's behalf. Four behaviour categories are targeted: boosting, buying, selling or transferring accounts, intentional deranking, and climbing via smurf accounts. These four are not isolated; they form a supply chain.

Picture that chain as a market. On the supply end sit high-skill players — Riot's blanket term — paid to play on other people's accounts. In the middle sits the account market: buying, selling, transferring, where a high-ranked account is priced like merchandise. On the demand end sits the payer, usually someone who wants a rank badge without putting in a thousand hours of climbing. And at the fringe sit auxiliary behaviours: intentional deranking to manufacture skill gaps, or using a smurf to climb alongside a main account.

This is where I have to say something analysts usually dodge. The transfer market is not a chessboard, it is a poker table — people bet money using reputation. I wrote that for football, and it holds exactly here: what trades in this rank market is not points, it is prestige. A Radiant account does not sell because it is strong; it sells because the buyer believes that position says something about them.

What Riot does with Anti-Boost, at its deepest layer, is intervene in that belief. A ranked ladder only has value if players believe rank reflects skill. The moment that belief collapses, rank is decorative — and the entire scouting pipeline, from solo queue to academy to professional roster, loses its raw material.

This is what outsiders miss. People think the boosting market is about a few stray accounts. It is not. It is about the whole talent-production chain. When an academy anywhere in the world hunts for a young marksman, they start from the solo-queue ladder. If that ladder is diluted, every downstream step is built on dirty data.

The penalty ladder: four tiers, and one nobody noticed

Riot designed a four-tier ladder. Tier one: on detection, ranked points and rewards earned through cheating are cancelled, the account is returned to its original rank, with a temporary suspension. Tier two: repeat offences escalate ban duration. Tier three: account buying/selling or intentional deranking can lead to a permanent ban. Tier four — the tier I care about — is joint liability: the booster's main account, and even teammates who frequently queue with them, may also be actioned.

There is a hidden inference in the mere existence of tier two. If Riot designed an escalation ladder, Riot assumes a recidivism rate high enough to require one. In a system where recidivism is near zero, escalating penalties are decorative text. Its existence signals something: Riot is dealing with a population that learns fast and returns fast.

Losers tell you about Zahavi, winners tell you about the number. Here the number is 296,416. But that number carries a structural problem I have not seen anyone flag: it is a cumulative total across two titles with entirely different boosting economies, unsplit by region, unsplit by title, unsplit by time period.

VALORANT is a 5v5 tactical shooter. League of Legends is a MOBA. Rank-inflation pressure differs fundamentally between them. In VALORANT, one high-skill shooter can nearly carry a match alone, so the value of hiring a good player is higher and boosting demand clusters at mid-to-high ranks. In League of Legends, role structure is more distributed, so boosting looks more like carrying a whole stretch than carrying a single match. Pooling both into one number is a communications choice, not an analytical one.

And there is another detail: no regional split. This is the most regrettable omission, because boosting demand is very unevenly distributed by region. In markets where accounts trade openly, where rank prestige converts into money or into streaming career opportunity, boosting demand runs far higher. A single global figure erases any ability to compare regions — and erases any ability to know where this policy is working and where it is failing.

Germans think they have drawn the map; I only need to see where their finger lands on the paper. Riot handed over a beautiful map: four violation types, four penalty tiers, one big number. I do not read the map. I read where the finger lands — and that finger lands on the "frequently queue with" clause.

Let me be precise about why that clause is a risk zone. When you solo-queue at a high rank, you encounter the same pool of players repeatedly. That is structural to every high-rank ladder: small population, overlapping play hours, limited server regions. Two random players can become "frequently queuing together" simply because they play the same evening hours, on the same server, at the same rank. Riot publishes no threshold defining "frequently." No appeal mechanism for this group. No criteria distinguishing a healthy climbing duo from a duo where one side is boosting.

This is where I reach for the word I use about systems designed with too much ambition: perfect. An automated system, scanning hundreds of thousands of accounts, described in the language of tidiness — perfect. But tidiness has a price. The price is that a group of players who violated nothing can sit inside a risk zone with no way to know in advance.

The biggest blind spot: an intent-based standard

Riot drew a subtle line, and I have to concede they did this more intelligently than most publishers. They do not ban alt accounts. Self-created, self-operated alts are normal activity. Anti-Boost targets intent to manipulate rank, not the existence of alts.

This is a narrow, deliberate standard, and as rule design it is correct. Ban alts and you ban behaviour most of the community considers normal — and you generate unnecessary backlash. Allow alts but target intent, and you keep public goodwill while still hitting the real behaviour.

But an intent-based standard has an operational problem no design can erase: it is hard to apply transparently. A bright-line rule — ban all alts — is easy to explain, easy to verify, easy to appeal. An intent-based rule must infer intent from behaviour, and inference always carries error. That error, multiplied across hundreds of thousands of accounts, produces a large volume of grey-zone cases.

And this is the point I want to stress most in this entire piece: Riot is simultaneously the detector, the prosecutor and the judge, and no independent appeals body is described anywhere in the documentation. All governance authority sits with the publisher. That does not mean Riot acts wrongly. It means that when Riot acts wrongly, there is no external mechanism to correct it.

In football, when a player is suspended unjustly, there are sports courts, investigative journalism, federations at multiple levels. Here, that chain collapses into one. Who says esports is a sport? It is a stock market with no holidays — and the exchange is also the regulator, the court, and the auditor.

A self-reported number and the problem of unaudited evidence

There is a detail in this data that analysts should handle more carefully: 296,416 is self-reported by Riot, with no independent audit. That does not make it false. It means we are reading a self-graded report card.

Worse analytically: it is a cumulative total with no prior-period baseline. A total does not establish a trend. When the source article interprets this figure as "Riot is tightening its grip," that is the writer's inference, not the data. To claim tightening, you need at least two time points. Here there is one.

I learned this lesson the hard way. In 2026 I wrote that Guangzhou R&F had to sell Eran Zahavi immediately, arguing their defence was fragile and masked by his 27 goals. They finished fifth and conceded 46 — the numbers sided with me. But I also remember that because one number was right, I nearly pushed the argument further than the data allowed. A correct number is not the same as a correct conclusion about a trend.

Same here. 296,416 actioned accounts is a fact. "Riot is tightening" is a reading. The two must be separated.

There is one more thing about this number: it counts accounts actioned, not accounts correctly actioned. In any automated enforcement system, those two figures are not equal. Read this number as an effectiveness metric and you are reading an activity indicator as a quality indicator. That is the single most common analytical error when reading any organisation's enforcement report, from sports federations to tax authorities.

72 hours, and the lesson of reading what is not in the data

There is one more thing in this documentation I want to address, but from a different angle: the expansion plan. Riot says it will keep expanding Anti-Boost and add match-level detection of boosting signals. They admit current methods are imperfect.

I read that and remember March 2026. The pandemic froze every pitch, my company cut 40 percent of staff, and I lost my job just as my name was rising. Within 72 hours I built a livestream room at home and started a Liverpool series. Those hungry, sleepless 72 hours taught me: a pitch is also a form of epidemic front line. That sounds off-topic in a piece about an anti-boosting algorithm, but the principle is identical: when you lose the thing you leaned on, you must build the new thing before you have time to understand what you lost.

Riot is in exactly that position, at larger scale. They admit detection needs improvement, meaning they know this race has no finish line. Boosters adapt. Detection chases. Both sides learn. That is the nature of every anti-fraud war, from doping in sport to tax evasion in economics. Nobody wins forever; only whoever holds the equilibrium longer.

There is one adaptation direction the documentation omits, and I believe it is the next chapter. As account-level and match-level detection thickens, boosting flows will migrate to harder-to-detect channels: off-platform communication, organised deranking rings, or exchanges that leave no trace in server systems. That is a general law: the greater the pressure, the deeper the behaviour burrows. There are no exceptions in the anti-fraud history of any industry.

Where I could be wrong here

I need to state this clearly, because a critique without self-examination is just noise.

First, Riot may already have an internal threshold for the "frequently queue with" clause — they simply do not publish it. If that threshold is high enough, wrongful-enforcement risk drops sharply. I have no data to conclude, and this is where I could be wrong.

Second, the wrongful-enforcement risk may be smaller than I am inflating. In practice, joint-liability cases may all involve duos with determinate relationships, not strangers overlapping hours. If so, the design is more reasonable than I describe.

Third, I am reading a summary document. The original may contain appeals-mechanism details this summary skips. I judge on what I have, not on what exists.

What deserves restating

What Riot has achieved, in the end, exceeds what it has not. They publicly published a clear violation taxonomy. They set an intent-based standard, subtler than the crude alt-account bans other publishers still use. They disclosed enforcement figures — an act of transparency not every publisher dares, because disclosure means placing yourself in the crosshairs.

But there is one question this documentation does not answer, and I believe it will shape the next twelve months: what happens when a famous wrongful-enforcement case appears?

Not if. When. Every automated enforcement system at the scale of hundreds of thousands will produce at least one such case. The question is not whether the system errs, but how it corrects and how fast.

My prediction, and this one is verifiable: within twelve months, at least one player actioned under joint liability without directly violating anything will surface publicly on community forums, loud enough that Riot must respond. If that does not happen, I misjudged the risk of the "frequently queue with" clause. If it does, we will learn something more important than the 296,416 figure: that in this ecosystem there is no court other than the publisher — and that is esports' real problem, not a few boosting accounts.

The crowd fears being wrong so it picks the strong team; I pick the right one — and only I know it. Here I pick no team at all. I just read the clause nobody wants to reread.

Cầu thủ liên quan