Rabbit Ears in the Esports Feed: How a Taxonomy Error Is Eroding Gaming-Industry Data

On Tuesday night I sat in my small apartment in Incheon and opened the international feed, as I do every evening. Between dozens of headlines about group stages, suspensions and publishers' apology letters, one image sat squarely in the middle. In it: a young woman with tall white rabbit ears, a navy sailor outfit, a plastic sword in her hand, her eyes mischievous as though she had just thought of something.

Underneath the photo, the taxonomy tag read two words: Esports.

I sat still longer than a photo deserves. The photo is beautiful, and I will get to that beauty later. But something did not fit. The caption spoke of an "easily recognisable design", of the ability to "transform through many outfits", of a character that is "hard to ignore". Those words belong to an entirely different economy. And they had just been filed in the same drawer as matches deciding qualification for Worlds.

The character is Shimakaze, a Sakura Empire destroyer from Azur Lane, a mobile gacha game developed by Manjuu and Yongshi. Azur Lane has no professional circuit at a scale comparable to League of Legends, DOTA 2, CS2 or Valorant. It has no franchise system, no slot purchases, no regional qualifiers, no weekly power rankings. Its content cycle runs on new character and skin releases, not on balance patches.

Yet the article about it sits in an esports vertical. It is signed by Tuan Hung. Right beside it, in the related-links block, are headlines that genuinely belong to esports: PUBG Asia Stars, Vietnamese player Himass facing a possible suspension, friction between KRAFTON and organisers, and another line about the CEO of Box Vietnam being wanted over copyright infringement.

I have seen this structure hundreds of times in nine years on the beat. An outlet runs two content streams in parallel: a hard stream of results, transfers and discipline; a soft stream of cosplay, characters and fan culture. Both share the same tag system, the same database, the same label. And here the label was wrong.

My concern is not that the cosplay article exists. Fans making content about characters they love is healthy, and an outlet earning traffic from it is ordinary. My concern is that a subject with no competitive dimension whatsoever is being counted inside esports content volume.

Start where the data actually lives. An esports feed survives on two things: frequency and density. Frequency is posts per day. Density is the share of posts carrying competitive signal — match results, roster changes, contract disputes, officiating complaints. When a cosplay post is tagged esports, frequency rises and density falls. Density is what sustains a sports news product.

I once sat in a meeting room at Inven Global in late 2026 as the analytics team presented a breakdown of posts by tag. Nobody in the room was surprised to see the "esports" bucket ballooning. Everyone knew why. Nobody fixed it, because fixing it meant rewriting tagging rules — work that brings no traffic, only accuracy.

What is called the "meta" of Azur Lane is not a meta. It is commercial design. Shimakaze is popular because her silhouette reads instantly and because she can wear a great many outfits. In the language of gacha, that is skin monetisation. In the language of esports, that capacity does not exist. No tier list is upended when a character launches. No win rate is recorded. No team has to change its draft.

The real value of this cosplay product flows through a different chain: character design → fan expression → cross-community pollination. This is the flywheel of a gacha IP. It spins on attachment, not on results. The cosplayer in this set is known as Tieshou Jiaoshou, and the description praises her ability to channel the character's mood — specifically a "playful spirit" and a feeling that she is "closer to the in-game version". Those phrases belong to aesthetic criticism. They carry no data. They cannot be verified. They are promotion written in the form of a review.

I say that to name it accurately, not to sneer. A promotional piece has its own value. But when it flows into a data pipeline used to measure the health of a competitive discipline, it becomes noise.

The consequences are not abstract. A sponsor reads a report on esports content growth; the curve points up; money flows into a vertical where part of the output never discussed competition. An editor reads a report on trending topics; cosplay tops the list; more cosplay is commissioned; the ratio shifts further. Six months later nobody remembers why the esports section is full of photographs.

From Incheon I see this more clearly in smaller markets. In Korea, an Azur Lane cosplay piece stays inside the mobile-gaming or anime vertical, because newsrooms here have the headcount to separate lanes. I have watched major Korean outlets maintain two separate tagging systems: one for esports, one for game culture. A cosplay set never leaks into the results feed, because editors know the two audiences arrive with different needs. Merging them saves a few hours of operations and costs years of credibility.

In Vietnam, where a single outlet runs many verticals with a thin team, lane separation is routinely skipped. "Esports" becomes a bag for anything game-adjacent.

Twelve thousand reads in six hours — I once hit that figure with an exclusive about Park Seung-min, a nineteen-year-old amateur I found in a PC bang in Incheon. I held the information for weeks, waiting for the right moment to publish. That discipline — waiting for the right beat — is something an automated tagging system will never have. It only knows how to push everything out immediately and slap on it whichever label pulls hardest.

I know what it is to work with dirty data. In 2026, to write that exclusive, I hand-recorded the statistics of forty collegiate matches because no database covered that tier. A 72% lane win rate, an undefeated record in game one — I counted every number myself. Had someone handed me a file mixing that mid laner's stats with the skin list of a gacha game, I would have lost another two weeks just cleaning it.

There is another layer few notice. The cosplay set exists thanks to publishers' tolerance policies for fan works. That is a separate legal layer, distinct from tournament rulebooks. The headline about the Box Vietnam CEO wanted over copyright infringement sitting beside this photo set shows both belong to the story of intellectual property in gaming. Only one of them is esports news.

The fan-content flywheel runs like this: an appealing character design is released; the community receives it; a cosplayer translates it into a real image; that image reaches people who have never opened the game; some of them download it; a smaller share pay for skins. No viewer sits watching two teams fight anywhere in that chain. That is precisely the point.

I remember Worlds 2026, held right here in Korea. I was seventeen, university entrance exams looming, and I did not miss a single match. Gen.G — the defending champion — won one group-stage game and went out. No LCK team reached the semifinals. I cried from the shock of my home region's collapse. Then IG of the LPL swept FNC 3-0 in the final, with TheShy on Aatrox landing nine multi-target sword swings. That night I wrote "The LCK fell silent, the LPL roared", 1,400 words, shared more than 500 times.

Those nine sword swings can be counted. They can be argued about. They can be proven wrong. That is what separates sports analysis from cosplay promotion.

Here I have to check myself.

There is another reading, and it is not naive. Tagging a cosplay article as esports may well be a marketing machine working exactly as designed. A cosplay post usually outdraws a tactics breakdown. Both sit in the same advertising folder. Both bring in money. The wrong label is a commercial choice dressed up as an operational error.

And while I sit here complaining about information density, I should remember that the first light of my career came from an analysis piece blended with free verse. In 2026, aged sixteen, I wrote "Karma's Dance" about Bdd landing twelve bindings in the deciding game of the LCK Summer final at Incheon Samsan. No editor checked my numbers. It earned 210 upvotes and a column invitation. Had anyone told me then that poetry did not belong in a sports feed, I might never have written again.

The difference is this: my piece told the story of a real match, with a result, an opponent, a consequence. The cosplay piece tells the story of a character who competes nowhere. Both carry emotion. Only one of them can be proven wrong.

The Longzhu shock is not meant to be erased, but to be sung. I still hold that sentence today, and it does not contradict what I am writing. The memory of a historic shock deserves to be sung. But it must be sung in the right temple.

I will not ask anyone to stop making cosplay. I will not ask any outlet to stop publishing it. I ask only for one small thing: call it by the name of the place it belongs to. A "Gaming and Fan Culture" vertical can hold both rabbit ears and a standings table. An "Esports" vertical cannot hold both and still mean anything.

People look at the scoreboard; I look at the cracks in the tactics.

And sometimes the crack is not in the tactics. It is in the label above the headline.

Every match is a draft, and only real writers dare keep writing it.