From a 100% Guarantee to a 9% Reality: The Arithmetic of Player Trust in KRAFTON's Four Crises

In 2026, a sentence appeared in the in-game store of PUBG: Battlegrounds, written in the language of a contract: if a player failed to receive the Set Blueprint four times in a row, the fifth pull would be guaranteed at 100 percent. Store text is not marketing poetry. It is a promise attached to a price. A player opening a paid bundle is really buying a probability, and when the language of that probability is wrong, the transaction stops being a product sale and becomes a misinformed transaction.

At my desk in Bengaluru, a large part of my routine is building pricing tables for esports markets — reconciling odds, drop rates and prize pools for the Indian audience. That work builds a habit: whenever I see a declared number, I record it in a form someone else can rerun and verify later. So the store line did not become a story to me. It became a test case. I wanted to lay it against the spending of every player who bought a bundle and measure the exact gap between the promised guarantee and the drop that actually occurred. The answer arrived later, and it sits far from the promise.

The problem was hard to raise publicly at first because the promise was public while the reality was private experience. Several players claimed they had bought more than five Premium Bundles and still had not received the Set Blueprint. Some said five, some six, some seven — the numbers were scattered, but the pattern was identical: the guarantee line was not being honoured. After a few hundred complaints accumulated, the team acknowledged that the Premium Bundle was in fact outside that pity mechanism, and that some in-game information displayed was inaccurate. That acknowledgement is the heart of the case, because the text the player read before spending is being called wrong by the publisher.

KRAFTON then corrected the information, apologised to players and announced a compensation plan. Even so, some who had spent heavily on bundles demanded cash refunds. In June 2026, a Yonhap report revealed that South Korea's Fair Trade Commission (FTC) had fined KRAFTON 2.5 million won for providing false item drop rate information in PUBG: Battlegrounds. According to the FTC, the company told players they would certainly receive an item after four failed attempts, while the actual rate was set at 9 percent. Some other items' rates were also displayed inaccurately. Because the company voluntarily corrected the information and paid compensation, the action was limited to a fine.

That is where the real question sits. This is not a temporary lapse by one company. It is a failure of disclosure infrastructure in a market where the "price" of a product is a probability, yet there is no mandatory public audit of that probability. The numbers printed in the in-game store are, in effect, price tags. And when a wrong price tag spreads across hundreds of thousands of transactions, it is no longer an individual error — it is a market distortion.

The NewJeans collaboration: two paid products selling one probability

The 2026 collaboration was an agreement between PUBG and the girl group NewJeans, placing two clearly paid products in the store — the PUBG X NewJeans Loot Pack and the PUBG X NewJeans Premium Bundle. Opening these gave players collectibles and character customisation inspired by NewJeans. Customisation is not merely recolouring; it is identity construction, which is why player emotion attaches to it. Where emotion attaches, price sensitivity falls.

That falling sensitivity is convenient for a publisher. A player buying seven or eight bundles has already suspended reasoning and committed. Translated into personal spending, if the promised five bundles had been real, every player's maximum cost would have been capped at a known ceiling. Breaking that promise removes the ceiling, and spending has no roof. This is where loot box design is strategically risky: it converts a collector's desire into cash flow, and that conversion rests entirely on the declared probability.

Guarantee versus independent drop rate: two different economies

A pity mechanism and an independent drop rate are not the same thing, and this distinction is the largest factual gap here. In a pity mechanism, every failed attempt raises the probability of the next; eventually, at a defined step, probability reaches 100 percent. In an independent drop rate, each attempt is separate, and the next attempt's probability is unaffected by previous failures. The two produce entirely different price structures.

Under the declared rule, one Set Blueprint is guaranteed within every five bundles, so the maximum cost of obtaining the set is five bundles. But if the actual rate is 9 percent per pull and each pull is independent, the probability of getting nothing after five pulls is 0.91 to the power of five — roughly 62.4 percent. In plain terms: where the declared system offered 100 percent certainty, under the real system nearly three out of five players who open five bundles walk away empty.

The expected number of pulls to a first success at a 9 percent rate is one divided by 0.09, or about 11.1. So the average player must open roughly 11 bundles where the promise was five. Cost rises by 2.2 times. To reach a 95 percent confidence level — where only a 5 percent chance remains of getting nothing — you would need about 32 bundles, which is 6.4 times the promised five. Both figures directly contradict the single line in the in-game store, and this is the factual gap for which the Korean regulator imposed a fine.

In Bengaluru I built a model. The first thing it killed was blind faith in declared numbers. That model had worked on home bias in football; the same principle applies here, with the publisher occupying the home team's seat. Without measuring the gap between declared and real probability, every decision rests on a false foundation.

The compensation arithmetic: refunds and in-game money are different things

The next stage was financial. KRAFTON refunded roughly 1.1 billion won to about 380,000 buyers and provided compensation worth about 9.8 billion won in in-game currency. At first glance these are large numbers. Divide them, and the picture changes: 1.1 billion won divided by 380,000 is about 2,895 won per buyer. The cash refund may therefore be less than the price of a single bundle that a player purchased.

The second number matters more. The 9.8 billion won was not paid in cash but in in-game currency, whose value is set by the same institution whose declared rate was wrong. If a player spends that currency on a new loot product, the compensation partly flows back toward the loss. This is why many insisted on cash refunds; it is not a grievance about price, it is a question about the independence of compensation.

Here my budgeting habit applies. Compensation should be seen in two layers rather than as one figure: verifiable cash, and institution-controlled credit. The first can leave the market; the second circulates inside it. When a firm corrects an error with credit, it is protecting its future sales in that same market.

A 2.5 million won fine against 1.1 billion won in refunds

Per Yonhap's June 2026 report, the FTC fine stood at 2.5 million won. The figure matters as policy because it establishes that false drop rate information is a punishable violation. On scale, however, the asymmetry is visible: 1.1 billion won refunded in cash, 9.8 billion won in in-game compensation, and a 2.5 million won fine.

Part of the explanation is voluntary correction; the regulator itself noted that because the company corrected the information and compensated players on its own initiative, the action was limited to a fine. The goal is not to destroy a company but to change behaviour. Yet from a market perspective the important thing is the signal: if misinformation is this cheap in the consumer market, what is the financial distance, for a company, between declaring correctly and declaring wrongly? No regulator answers that. Market repetition does.

The NewJeans face skin: design responsibility or user conduct

The same collaboration produced a second controversy in 2026 with an entirely different nature. According to Korean media, some players used face skins of NewJeans members to create and share obscene and harassing images and videos, combining the faces with revealing outfits. The case became more sensitive because Haerin and Hyein were not yet adults at the time.

KRAFTON and ADOR, NewJeans' management company, said they would act against uses that conflicted with the collaboration's original purpose. KRAFTON then restricted combining NewJeans face skins with certain in-game outfits. The decision generated fresh debate in the community: one faction argued the problem lay not only in user conduct but in PUBG's own character design and customisation system.

There is a clean way to test these two arguments. Moderation is a reactive system — it responds after an event. Design is a pre-emptive system — it creates or reduces the likelihood of an event. A face skin is an identity surface, and the outfit system grants independent mixing freedom; the probability of harassing use at that intersection can be tested with system rules, not only with user behaviour. The 2026 restriction was a moderation fix. A design-level fix remains an open question.

The China question: a 2026 IPO filing reviving an old debate

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 Tencent-operated Peacekeeper Elite and received service fees. That disclosure revived older questions.

The background is clear. PUBG Mobile shut down in China in May 2026 over licensing problems. Tencent then moved players to Peacekeeper Elite, a game with many similarities to PUBG Mobile but altered gameplay, visuals and content. KRAFTON had previously asserted the two were entirely separate products. One view held that Peacekeeper Elite had been altered enough to meet licensing requirements; another asked whether it was a route to keep PUBG alive in the Chinese market.

What matters to me here is the disclosure timeline, not national sentiment. When a service relationship first becomes public in a financial document such as an IPO filing, it becomes commercial fact. The distance between that fact and the earlier "separate products" statement is the question. Investors deserved clarity at the time of the licensing dispute, not before listing on the stock market.

PUBG Asia Stars 2026: an evidentiary crisis built around one decision

The most recent chapter concerns esports integrity. At PUBG Asia Stars 2026, two Vietnamese players were involved — Himass of Anyone's Legend and TanVuu of The Expendables. On 23 September, KRAFTON published its investigation findings and stated that both had used outside information, particularly another player's livestream, to form judgments and build tactics inside the match.

KRAFTON concluded this was stream sniping, violating PUBG: Battlegrounds operating policy and the conduct regulations for professional players. As a result, both accounts were permanently blocked and both were banned from any official PUBG Esports competition organised or approved by KRAFTON, including PGC, PGS and PNC.

The incident began with an accusation by Korean streamer Soopi of Gen.G, who claimed the two Vietnamese players were following opponents' livestreams. Organisers then removed Himass and TanVuu from the rest of the tournament, adjusted scores, increased livestream delay, cancelled the third day of competition and split the prize money. After the investigation, KRAFTON said no additional violations were found in other cases.

I have spent years watching matches with timestamp tables. That habit has taught me one thing: the credibility of a ruling is determined by the distance between how a dispute is raised and how the evidence is kept. Here the accusation was clear and the decision was fast, but how reproducible the evidence chain was has not been fully opened to the community. If a spectator cannot verify it themselves, the decision remains weak even when correct.

Why the two cases moved at such different speeds

Now the observation that stands out most across this entire history and is discussed least. In one case, users' money was directly deducted on the basis of false information — 380,000 people, billions of won — and the process was corrected slowly over months. In the other, a permanent ban arrived by 23 September, a competition day was cancelled, scores changed and prize money redistributed.

Consumer protection moved with evolutionary, measured speed; competitive integrity moved instantly and completely. Asking which matters more is the wrong question. The right one is whether both areas should carry the same evidentiary and explanatory obligations. If an institution holds full power for rapid sanctions but moves slowly on wrong price information, that is not moral failure — it is an incentive structure.

This is my strongest objection, and it is not support for either side. Stream sniping is a real problem: in a live streaming environment, inferring an opponent's position is technically possible, and it damages not only the rules but the value of the spectator's ticket. But the process that produces such a ruling has its own price: the universality of evidence. If a player does not know which information, at what delay, on what seconds of evidence will lead to a ban, the path of learning from sanctions closes. Only fear remains.

I return to that line — set pieces are not luck; they are rehearsed mispricing. Stream sniping enforcement is not luck either; it too is a matter of rehearsed, published, verifiable protocol. Otherwise every decision weakens, even the correct ones.

Community reaction: 4.1 million signatures and a squad-level fracture

After the ruling, the Vietnamese PUBG community's reaction was extensive. More than 4.1 million signatures were collected demanding fairness for Himass and TanVuu. Among content creators and streamers, Độ Mixi, PewPew, Rambo, DjChip and Ngân Sát Thủ reacted in various ways — some protested, some removed PUBG. GAM x The Expendables demanded a clear explanation from organisers, while Anyone's Legend stepped forward to defend the players' right to be heard.

The debate then stopped being confined to one question — did they break the rules or not. It spread in three directions: whether the severity of the punishment was right, how organisers handled the incident, and whether the same standard is applied to all participants. That third question is the most dangerous, because it is not about a specific incident but about the credibility of the rules.

From an incident at a showmatch, the story reached a national scale where community trust in the operator is directly questioned. Diplomatic statements do not work at that point. Only public, reproducible explanation does.

What to watch in the next round

In the loot box case, a regulator established one sentence: false declared drop rates are punishable. That is a precedent, and its effect will spread beyond PUBG — wherever probability-based products are sold, the language of the declaration will be tested. A studio assuming this is a Korean matter may be leaving other consumer-protection authorities out of its calculation.

On esports integrity, what is needed is not a fine but a protocol. If the methods used to detect stream sniping — timestamp delay, view records, IP or pattern analysis — remain secret, every future ruling will sound like an accusation. I want published operating thresholds: what is visible at what delay, how much of a pattern triggers investigation, and where punishment stops.

The biggest question is economic. When the price of an in-game item is written in the language of probability, that writing is part of the contract. I will read every future collaboration and loot product declaration with these questions: is this a drop or a pity? Is the probability independent per pull? And if someone reruns that number, will the same answer come back? A publisher that answers with no but compensates in credit leaves the account permanently incomplete — because money kept outside the market and money cycled inside it are never the same.