Four KRAFTON Incidents, One Disease: Rules in Public, Decisions in the Dark

September 23, 2026, 11:40 pm. I was on a balcony in Shenzhen, watching the restarted stream of PUBG Asia Stars. One small thing caught my eye: the live delay had been extended. In an event where viewers are used to a two-to-three-second sync, there was now roughly a full minute of separation. There are few simpler fixes for stream sniping. But delay cannot cover the problem it was built to hide — the decision process.

The same day, KRAFTON announced that two Vietnamese players, Himass (Anyone's Legend) and TanVuu (The Expendables), were banned permanently. Not just from PUBG Asia Stars. From PGC, PGS, PNC — every official PUBG Esports event KRAFTON organises or sanctions.

Within hours, the Vietnamese PUBG community erupted. More than 4.1 million signatures were collected demanding fairness. Do Mixi, PewPew, Rambo, DjChip and Ngan Sat Thu protested publicly, some uninstalled the game. GAM x The Expendables demanded a written explanation. Anyone's Legend defended their player's right to be heard.

Two questions sat in the middle of the argument. One was answered repeatedly. The other has not been answered once. First: did Himass and TanVuu actually break the rules? Second: where is the proof that the same standard was applied to everyone the same way?

This piece works on the second question. The Himass-TanVuu case is not an isolated accident. It is the latest instalment of a four-year pattern. And that pattern belongs not to any player's character, but to a company's information management.

Context: One company, four kinds of information crisis

The incidents look different. The structure is identical.

First, 2026. In the PUBG and NewJeans collaboration, KRAFTON sold two paid products: the PUBG X NewJeans Loot Pack and the PUBG X NewJeans Premium Bundle. Players could open them to receive collectible items and NewJeans-inspired character customisation. The problem appeared in the odds disclosure. KRAFTON announced that if a player did not receive the set blueprint after four consecutive opens, the fifth would be guaranteed at 100%. In practice, some players bought more than five Premium Bundles and still did not get the item. Under community pressure, the PUBG team admitted the Premium Bundle was not actually covered by that anti-bad-luck mechanism, and that some information displayed in-game was inaccurate. KRAFTON later corrected the text, apologised and issued compensation. Even so, players who had spent large sums demanded cash refunds.

Second, June 2026. Yonhap reported that South Korea's Fair Trade Commission fined KRAFTON 2.5 million won for providing misleading information about item drop rates in PUBG: Battlegrounds. According to the FTC, KRAFTON told players they would certainly receive the item after four failed attempts, while the actual rate was set at only 9%. The company was also found to have given inaccurate information on the odds of other items. KRAFTON refunded roughly 1.1 billion won to around 380,000 buyers and paid out about 9.8 billion won in in-game compensation. Because the company corrected the information on its own initiative and carried out compensation, the penalty was limited to a fine.

Third, another layer of the same NewJeans collaboration. In 2026, Korean media reported that some players had used face skins of NewJeans members to create and share sexually harassing images and videos by combining the faces with revealing outfits. The case drew more attention because Haerin and Hyein were not yet adults at the time. KRAFTON and ADOR, the group's management company, said they would act against uses of the items outside the collaboration's original purpose. KRAFTON then restricted combinations of NewJeans face skins with certain outfits. That decision also split the community.

Fourth, China. In 2026, ZDNet Korea reported that in its securities registration filing ahead of its IPO, KRAFTON publicly disclosed for the first time that it provided technical services to Peacekeeper Elite, operated by Tencent in China, and received service fees. Earlier, in May 2026, PUBG Mobile shut down in China after licensing problems, and Tencent moved players to Peacekeeper Elite, a game with many similarities but changed gameplay, visuals and content. KRAFTON has repeatedly said the two are separate products. Some argue Peacekeeper Elite was altered enough to meet licensing requirements; others ask whether this was a way to keep PUBG alive in the Chinese market.

Four incidents, four kinds of question. One disease.

The real analysis: not false information, but insufficient information

I live in Shanghai. In May 2026 I was a student in this city. What I saw then was that the people posting most about the shutdown were not players — they were traders, people reading registration filings, and content creators. The same thing happened when the 2026 documents became public. Nobody was surprised. An assumption simply became a record.

That is where the structure shows itself. KRAFTON does not have a habit of lying. Its habit is telling the truth in a way that does not help the user make a decision. That is not an accident. It is design.

The maths of four-then-five

A drop rate and a pity system are two different objects. A drop rate describes the present: your probability on each open. A pity system is a contract about the future: after N failures, you receive it with certainty. Both can coexist, but the existence of one does not prove the existence of the other.

Had KRAFTON said 'this bundle has a 9% drop rate and no guarantee', players would have bought or not bought — the decision would have been theirs. What was said instead was a story structure: 'fail four times and the fifth is yours.' And it did not apply to one paid bundle at all, while in another tier of the same purchase the real rate was 9%. The customer held certainty in their head and probability in their hands. In gacha economics that is the most profitable gap of all: sell confidence, deliver odds.

One line I use for every esports calculation applies here. The cosmetic market is not a spreadsheet. It is a market of stories, and the price of a story has to be re-set every season. A price built from FOMO and collector identity is never something a firm wants to reduce to a probability figure, because the moment it is reduced, the margin enters the realm of human reasoning.

But run the numbers a different way and the picture changes.

Take 380,000 buyers and 1.1 billion won in cash refunds: the average buyer got back about 2,895 won. Take 9.8 billion won in in-game compensation: that is roughly 25,789 won of virtual value per buyer — value that never leaves the balance sheet as cash, only as data and liability. Then take the fine of 2.5 million won. Divide it across the affected users and you get about 6.6 won each.

Under seven won. Roughly six and a half won of penalty per affected user. This is not the only metric — Korean enforcement sets precedent, it does not bill per head. But the incentive story writes itself: cash refunds are real money, in-game compensation is paper money, and the fine is smaller than the fee. The path that should be most expensive — making a false promise — is the cheapest.

I have not seen this ratio laid out anywhere before: a compensation package in the millions-of-dollars scale, settled with a fine in the low-millions-of-won scale. The bigger the argument gets, the smaller the incentive rhythm becomes.

Face skins: a design problem relabelled as user misconduct

The NewJeans face skin incident clarifies the information-management thesis. Players combining character faces with outfits and sharing what they made is repugnant. There is no doubt about that. But the question does not stop there: why was the system built so that it was instantly possible?

PUBG customisation is a map. Every item is an open endpoint, and the number of combinations is effectively infinite. What the language of FOMO and item purchases calls creative freedom, the language of ethics calls moderation debt. The larger the state space of free combination, the more impossible moderation becomes. When KRAFTON built face skins, it turned an identity into a modular asset. Modularise it and the liability becomes modular too — belonging to the system, not only to the user.

Their fix? Ban certain combinations. That is not dishonest. It is incomplete. It fixes the output, not the grammar. Esports is a laboratory with no alibi: if the system itself can generate a near-infinite number of degrading combinations, the responsibility does not rest only with those who generated one.

The most uncomfortable part is the timing. The commercial logic of the collaboration was selling identity. Haerin and Hyein were minors then. The relationship between age and liability is never written plainly in commercial communication.

China: 'separate product' is a legal description, not an economic one

Some dismiss the China story as old news. I do not, because it sits at the exact centre of the pattern.

In May 2026, PUBG Mobile shut down in China over licensing. In 2026, KRAFTON's own filing revealed it provided technical services to Peacekeeper Elite and received fees. The space between those two facts is the lesson. 'Separate product' is a legally coherent description — it determines which file lands where. 'Technical service plus service fee' is an economic description — it says which hand the money moves from and to. Both can be true at once. But a reader who sees them together can see what fills the gap between them.

That gap is the central symbol of the pattern. Each KRAFTON statement, checked individually, holds. But what players and viewers wanted collectively — a map of the decision — has never been in any statement.

A point needs saying plainly: a public data chain should exist, where any player can check the odds themselves, verify them, reproduce them. The industry has not built that. Odds remain marketing copy, not an auditable record.

Stream sniping: fast punishment, slow process

Back to PUBG Asia Stars. It began when Korean streamer Soopi of Gen.G accused two Vietnamese players of watching opponents' livestreams. Organisers removed Himass and TanVuu from the rest of the event, adjusted scores, increased stream delay, cancelled the third day of competition and split the prize money evenly. Then, on September 23, KRAFTON's investigation concluded: both players used external information, including other players' livestreams, to form judgments and build strategy. That is stream sniping, violating the PUBG: Battlegrounds operating policy and the conduct rules for professional players. Both accounts were permanently locked and both were banned from all official events. No further violations were found.

The speed of the decision is admirable. But if speed is the only visible part of the process, that is the problem. What was the evidentiary standard — what share of the data, how many testimonies, which tools — was never stated. And third parties who broke no rule paid: cancelling day three reshaped the calendar, and splitting the prize money changed what the points meant. Teams who were not accused received no compensation at all.

There is a second layer. The accusation came from a competitor. Soopi may well be right — we do not know. But on a platform that speaks publicly about its anti-sniping tools and sanctions, verifying a complainant's conflict of interest should be a standard step. Either it did not exist, or it existed and is invisible.

The crisis completes itself here. Players and fans were not challenging the conclusion. They were challenging the process. And the language shifted: 'we apply the same standard to everyone' is not evidence. It is a claim.

What happens when information is withheld

Lay the four incidents side by side and a line appears. The 2026 odds statement, the 2026 fine, the 2026 face skin policy, the 2026 technical service disclosure, the 2026 investigation — in each, KRAFTON did exactly what remains legally defensible after being proven wrong. Correction, apology, compensation. The sequence is always the same.

Every hot take is a hypothesis wearing a jersey. I am putting mine down where it can be tested: in all four cases, the core failure is not falsehood. It is telling truths that do not inform. When there is a gap between two true statements, what the user sees is a decision blueprint cut down to details.

Which brings the question back around. Why not publish the process? If the process is sound, disclosure costs nothing. Disclosure becomes expensive only when the process is not reproducible.

Now the other side: I could be wrong

Let me write the strongest case against myself.

First, KRAFTON's conduct is actually above industry average. Most global publishers caught misstating gacha odds just edit the text. They do not apologise, and they do not refund. KRAFTON returned 1.1 billion won in real cash to 380,000 buyers, paid 9.8 billion won in in-game compensation, corrected the text and restricted skin combinations. The regulator showed leniency precisely because of that cooperation. What critics call a transparency collapse may in fact be a repair done well.

Second, part of the Vietnamese reaction is national narrative repricing. The accuser is Korean, the publisher is Korean, the sanctioned players are Vietnamese, and this is one of Asia's largest PUBG communities — a four-million-strong reaction carries its own emotional inflation. A Bangladesh-born writer watching from China has seen market arbitrage before, and exactly for that reason should not build a conclusion on it.

Third, the 6.6 won per user figure may be unfair. A Korean fine is a precedent, not a number. It changes future behaviour precisely because everyone sees it, and that is not measured per head. That is fair.

All three objections run into one test. If the process were solid, publishing it would cost the company nothing. If the standard were fixed, which document states it? If the apology were reproducible, it would be written somewhere within six months. The absence is the argument.

My prediction, with a date

With 60% confidence, I expect KRAFTON to build a public dispute-resolution protocol for PUBG Esports within the 2026-27 cycle, with a stated evidentiary standard and a hearing route, plus a verifiable audit page for item odds. If neither document exists before PGC 2026 ends, my reading is wrong, and I will write that.

Second: the reaction will be measured where 4.1 million signatures are not. Signatures do not move a balance sheet. Viewership moves it, stream retention moves it, seasonal item spend moves it. The next six months of PGS view numbers will show how much of this anger lives in a database and how much lives on a timeline.

One question deserves to be left standing, because it is the actual centre: if the rules are the same for everyone, why is the proof of following them not visible to everyone?