T1's File Before Worlds 2026: Oner at 5/6, Faker Near the Bottom, and the Data Lines Nobody Zooms In On

On 14 August 2026, a four-column spreadsheet arrived on my machine from Seoul. The sender was a data analyst who had worked with two LCK teams and asked to remain anonymous. The file had no title, no publishing entity, no timestamp in the footer. Four columns only: player name, fight participation rate, damage contribution share, and average gold difference per game.

The fifth row from the bottom in the first column carried a familiar name. Beside it, the notation 5/6. The third row from the bottom carried another familiar name, with a note in English scrawled in the comment field: near the bottom against eight teams.

Those two names were Mun Hyeon-jun, known as Oner, and Lee Sang-hyeok, known as Faker. Both play for T1. Both sit at the tail end of the 2026 season, with the domestic playoff bracket closed and the World Championship approaching.

I spent four days trying to identify who compiled the sheet. As of writing, I still do not know. That is why I do not call it evidence. I call it an unverified document, and what follows is everything I can do with it: read slowly, cross-reference, and mark clearly what is data, what is inference, and what is merely noise.

Why a spreadsheet never creates itself

No spreadsheet creates itself. It begins with a person choosing columns, choosing a sample, choosing a time window, and choosing how to label that column. I always read the footnote before I read the numbers at the top. Here, the footnote does not exist.

That does not make the numbers wrong. It makes them legally incomplete as information. A percentage without a sample, without a time frame, without a variable definition, is technically a bare number. It may be correct. It may also be the result of selecting the right six matches to produce the right conclusion.

My job is to read the smallest lines nobody bothers to zoom in on. The truth sits in the smallest lines nobody bothers to zoom in on. In this spreadsheet, the smallest line is the line with no words at all.

Context: a season told in two different halves

The 2026 League of Legends season has covered most of its distance. The document I hold references a six-team playoff, and elsewhere an eight-team sample. The two figures do not match, and that is the first thing I flagged.

Two possibilities. First, the league runs two distinct phases, a six-team knockout and an eight-team points stage, and the compiler merged both. Second, the compiler did not distinguish them, and the real sample is smaller or larger than stated.

With a six-team sample, cross-play is minimal. With eight, not much more. In both cases, a 5/6 ranking or a near-bottom placing across eight teams is highly sensitive: two bad series, or two series against strong opponents, can move a player several places.

I have worked with data from traditional sports where a season runs thirty-eight rounds. There, end-of-season ranking carries statistical meaning. Here, with six to eight teams in a short knockout, we are discussing a sample any serious analyst would flag before publishing.

The second context point is the patch. The document says gameplay shifted in many directions after updates, and that the jungle role still matters because junglers coordinate with supports and mid laners to control the map and pressurise side lanes. That is the entire patch content. No version number, no champion, no item, no mechanic.

That creates a serious methodological problem. Without knowing which patch, you cannot say whether it favours T1. Without knowing which patch, you cannot say whether it targeted T1's dominant style. Without a named patch, the sentence about changes is a framing device, not an argument.

The third context point is Worlds 2026. The document describes it as imminent. It does not name the edition, the start date, or the format. For a piece about the most important tournament of the year, the absence of all three is a signal about provenance: it was written from a feeling of proximity, not from a calendar.

Core: dissecting the three metrics

Fight participation

This is the share of team kills a player was present for. For a jungler it depends heavily on whether he moves to fights and whether his team creates fights where he already stands. A low figure can mean poor play, or a farm-and-objective style, or a team winning lanes so hard that jungle intervention is unnecessary. Ranking 5/6 across six teams means four junglers did better. In a six-team field, that gap is narrow. It cannot tell you whether Oner is being read by opponents or missing critical ganks. That requires pathing, timing, and gank-success data. The sheet has no such columns.

Damage contribution

Structurally, junglers deal less damage than mid and bottom laners in most metas, because their time goes to movement, vision and objectives. Comparing junglers to junglers is methodologically correct, and the document claims same-position comparison, which is a point in its favour. But it does not disclose the sample, the matches counted, or whether positional grouping was applied throughout. A low damage share for Oner has three explanations: fewer fights attended, tank/utility champion choices, or poor target selection inside fights. These lead to three different conclusions from one number.

Gold difference

This is net gold against the opposing player in the same role. For a jungler it is pathing-sensitive. A negative differential is not automatically bad; it can reflect deliberate resource sacrifice for a lane. But when a negative differential sits alongside low fight participation and low damage share, a shape emerges: less presence, less influence, less accumulation. That is when I began to believe there is a real signal, because all three metrics point the same way.

The undefined baseline

'Usual form' is never defined. From when? Which competitions? Does it include scrims? Does it exclude off-role games? Every decline claim depends on a baseline. If the baseline is a peak period, everything afterwards is decline. If it is a career average, the result can differ entirely. I read financial reports slower than others, because I read them twice. Here, the second reading of a four-column sheet gave me no new information, only new questions.

Contrarian: what T1's defenders are right about

A simultaneous dip in two long-tenured veterans is more likely a system-level signal than two separate individual collapses. Possible systemic causes: degraded scrim quality, coaching misreading the meta, a lost coordination piece, or plain late-season fatigue.

Oner has repeatedly been a criticism magnet. That is methodologically important: public reaction will exceed what the data permits. I saw the same pattern in Korean football years ago, when a defender was blamed for three seasons over two mistakes in one semi-final. His next-season numbers were better than the previous year, but his reputation was already frozen. A frozen reputation cannot be repaired with statistics.

A jungle-critical meta cuts both ways. High leverage amplifies good play and bad play. It means Oner's improvement is one of the clearest levers T1 can pull before Worlds, and it also means a coaching staff can reduce risk by selecting a structure less dependent on him.

Ranking 5/6 is not a verdict. In a six-team sample, the gap between fifth and third may be a few percentage points. Nobody publishes the gap, because gaps do not generate headlines. Rankings do. Finally, the missing data matters most: no injury data, no practice hours, no scrim count, no commercial obligation data. For a team with a globally branded player, those variables may outweigh every tactical variable.

Who benefits when someone else wins

Tier one: the compiler. A ranking creates order, and order serves a piece that needs evidence. This does not mean fabrication; fabrication is easily caught. Sample selection is far more common, legal, and hard to trace.

Tier two: platforms and the attention economy. A story about Faker declining outperforms a story about Faker being stable.

Tier three: personal brand versus competitive value. A related link in my file references a semiconductor chief executive meeting Faker alongside a phrase about internal power struggle at T1. Links are not article bodies, but they hint that Faker's commercial value may be decoupling from his competitive results. If so, a form dip does not reduce sponsorship value, and if sponsorship value holds, pressure to fix competitive problems does not rise in proportion. Money has no name, but contracts always do.

Tier four: calendar and national-team overlay. A continental multi-sport event with an esports programme fragments practice, rest and competition schedules. That is an invisible tax on performance.

Tier five: fans and pre-loaded expectations. A narrative built around a Worlds revival has two outcomes. If T1 revive, the story is confirmed. If not, the story has already installed a target for blame. Either way, the storyteller bears no risk. That is why I call it a story designed to be unfalsifiable.

Signals to track over six weeks

Patch identity and pro pick-ban data. T1 form across a full-season sample rather than a six-to-eight-team playoff slice. Any coaching or roster change, since no change means the diagnosis is unchanged, and change means the old diagnosis was wrong. Health and burnout signals, since wrist and mental fatigue appear in no ranking. International calendar overlap. High-tier commercial deals, which would confirm decoupling.

What I know after four days with four columns

There is a signal. It comes from a small sample, an unidentified source, and an unnamed season. If it is real, it does not say Oner or Faker are finished. It says T1 had an operational problem late in the season, and whether it can be fixed depends on whether the staff find a cause rather than a person to blame. No scandal starts with the janitor. It starts with the boss's signature. Here there is no scandal, only an unfootnoted sheet, a community looking for accountability, and a team entering the period when every old story gets retold. Every season ends, but a file does not. When Worlds 2026 closes, I will open this sheet again, read it a third time, and compare every line with the real result. That is the only way a bare number becomes a lesson instead of a rumour wearing a data label.