Karmine Corp Closed the LCP's Last Door: 28 Minutes, 18–3, and a Scorecard That Doesn't Reconcile

In any Worlds series, the highlight reel arrives first. On Thursday, 29 October 2026, what arrived first was a line: 28 minutes, 18–3.

I read that line three times. In League of Legends, 18–3 means one team beat the other by fifteen kills — a six-to-one ratio. Twenty-eight minutes means the gap did not take long to build. Put the two numbers side by side and the story looks complete: a laning-phase edge, converted at tempo, then the game.

The story looks complete. It is not. My work starts exactly where a story that looks complete drifts away from the record.

Karmine Corp beat CTBC Flying Oyster 2-0. With that result, the last LCP — League of Legends Championship Pacific — representative exited the Worlds 2026 Swiss stage. KC now plays the 2-2 pool decider on Saturday, 31 October 2026. The Grand Final is on Saturday, 14 November 2026.

This is a report on that match. It is not a report on the 18–3 you saw in your feed. It is a report on the record that exists — and on the record that does not.

The Swiss arithmetic, and a calendar audit

The Swiss stage runs on a simple premise: three wins and you are through, three losses and you are out. Those left in between meet in the 2-2 pool, where a single match decides whether the season ends tonight or continues next week. Round 4 meant that pool was nearly full, and every result was quietly rewriting somebody else's path. CTBC Flying Oyster was the last LCP entry in it.

One clarification matters for the regional story that will follow. What is documented is this: on 29 October 2026, the last LCP team left the tournament. What is not documented is how many LCP teams entered, what their aggregate Swiss record was, and how many reached 2-2. Without those three numbers, writing the end of a region is not a finding — it is an inference. I keep inferences in footnotes, not headlines.

The dates have their own audit. 29 October 2026 is a Thursday; the decider is Saturday, 31 October; the Grand Final is Saturday, 14 November 2026. I checked the weekdays against a calendar — all three reconcile. What is more interesting is the distance between them: fourteen days from decider to final. Fourteen days means two weeks without competitive matches, inside a scrim block, locked to a tournament patch. That gap is a gift to teams with a coaching structure and a trap for teams that improve match to match.

Now, what the record contains. Two game durations — 28 minutes and 37 minutes. The Game 1 kill line, 18–3, sourced to GosuGamers. And one sentence about side selection. That is the whole of it.

What it does not contain is a long list: no patch version, not a single champion pick, not a single ban, no Game 2 kill count, no gold differential, no dragon, herald or tower data, no player-level statistics. That gap is not a flaw in the report — it is a results piece, and expecting more from a results piece is not the reader's job. It is mine.

I will declare a bias here, because it is how I work. In 2026 I tagged 312 S.League and AFC Cup matches, built expected-goals and PPDA models, and filed a weekly chance-quality column. In 2026–22 I built a placement-adjusted BGMI rating across 640 maps — survival time, damage per minute, rotation efficiency, endgame conversion. And in 2026, when the circuit shut down, I hand-tagged a 1,180-match archive of PMIS 2026 and PMCO VODs. 1,180 empty matches taught me that absence is also a dataset. That lesson applies here: what was not recorded is still part of this series, and it has to be named.

18–3 is not a story about speed. It is a story about distribution.

The spreadsheet was the first witness, and it never blinked.

Game 1 produced 21 kills in total — 18 and 3. Twenty-one kills in 28 minutes is 0.75 kills per minute. Take any ordinary professional League game: 24 kills across 32 minutes is also 0.75 per minute. The pace of this game was not abnormal. What was abnormal was the split.

This is where the feed's story and the record's story diverge. Heard as '18–3 in 28 minutes,' it sounds like Karmine Corp played like a storm. But 0.75 kills per minute says this was not a storm — it was a monopoly. Not a bloodbath; a one-sided seizure. The difference matters: a monopoly involves less violence and more control.

A fifteen-kill gap is itself a technical statement. A gap of that size usually forms for one of two reasons, and the report does not separate them. First: KC converted every pick into an objective — a kill into a plate, a plate into a dragon, a dragon into a tower — and CTBC never got a window back. Second: CTBC contested fights it had no business contesting, and built the gap itself. The first credits KC's ceiling; the second credits CTBC's failure. They imply different things for the decider, and without objective data I will not choose between them.

Game 2 ran 37 minutes. Nine minutes longer — 32 percent longer than Game 1. The easiest and most wrong conclusion here is that a longer game must have been a closer game. Duration is not a proximity metric. A 37-minute game can be two entirely different objects: a 20–19 knife fight, or a 12–4 stomp that stalled at the nexus. Both write 37 minutes on the scorecard. The report gives no Game 2 kill line, so 'the second game was closer' is an inference with no support. That is the single largest information hole in this series, and it is exactly where the correlation-equals-causation error is most tempting.

Closing that hole takes two numbers, and both should be public: the Game 2 kill count and the gold differential at 20 minutes. The first tells you whether it was a fight or a stalled stomp; the second tells you who owned the game at 20. Without that pair, half the series analysis is conducted in the dark.

The sentence that collides with its own rule

I went looking for a scoreline and found a paper trail instead.

According to the report, CTBC had first choice for Game 1 and 'took first pick,' which placed Karmine Corp on the blue side. In the standard League draft, blue side holds pick one and red side holds the first ban. Which means the sentence collides with its own rule: if CTBC took first pick, CTBC was on blue, not KC.

There are only three routes out. One, the writer conflated side choice with pick order — the most common error, and the most innocent. Two, the event ran a non-standard draft order — a few tournaments have experimented, and if Worlds 2026 did, that should be on the record. Three, a straightforward reporting error. My lean is the first, my confidence is medium, and the result does not change: KC won the series, and that is documented.

But the consequence matters. A large share of post-match discourse assigns blame to the draft — who won it, whose composition fell off, who exploited blue side. For Game 1 of this series, that entire argument is unusable, because the sentence it rests on does not reconcile with the rules it invokes. I will not call such a conclusion false. I will call it unverified, and unverified conclusions do not become my headlines.

What survives about KC is thinner than it looks. They won two games of different shapes — a 28-minute fast one and a 37-minute slow one. I could call that multi-speed adaptability. The honest version is that this is a two-game sample, and two games do not make a team profile. Watching a team win fast games and declaring it master of two speeds is precisely the error I catch in others. I will call it a hypothesis, not a finding — and a hypothesis gets tested in the decider, not here.

What survives about the LCP is thinner still. On 29 October 2026, the last team went out. That is one date, one team, one series. It is not a regional verdict. If three LCP teams entered and all three fell at 2-2, the story is not an ending but a return from the doorstep. If only one entered, this is a story about a team, not a region. I do not have the denominator, so I will not write the easiest sentence available.

The more interesting question is who built the gap

Now the part I believe least, and suspect will be reported backwards. The 18–3 will be read as proof of KC's ceiling. I would argue the same number is equally consistent with CTBC's collapse, and the two carry different prices in the decider. If KC built the gap by converting every pick into an objective, the pattern travels, because patterns travel with teams. If CTBC built it by taking fights it should have declined, the next opponent will not hand over the same gift, and KC's real examination happens then.

The second thing I heard most after this match: the LCP is gone, the region is finished. I attach an asterisk, and I attach the condition that removes it. That verdict becomes real the day all LCP teams' pick-ban data, gold differentials and Swiss records are public — three numbers. Without them, 'the end of a region' is a feeling wearing a data costume. I will also say the feeling may not be wrong. It is simply not proven yet.

The third point is a weakness in my own work, and I will not hide it. I am writing from Singapore. This report is missing the thing that matters more than everything else — a floor source. No player, no stagehand, no local tournament organiser I could ask whether the hall was silent or loud in the 37th minute of Game 2. I archive the silence first, because silence is the first thing an institution edits. But here there is no record of the silence at all — not of the room, not of the rest of the scoreboard. Until someone who stood in that venue tells me what the 37th minute sounded like, this report is incomplete, and I will say so.

What to watch next

In Saturday's decider, I will watch one number: the length of Game 1. If KC closes inside 30 minutes again, the duration-versus-distribution reading repeats and becomes a pattern. If the same game stretches past 35, the two-speed hypothesis faces its first real test — and the question becomes whether KC has a way to win a slow game, or only a fast one.

And the number I will keep in the archive is not 18–3. It is the Game 2 kill line that was never published. Sometimes the number that is missing explains a series better than the number that exists. The next time someone sends you 18–3, ask what the second game finished on. If there is no answer, the story is still unfinished.