Six for Six: MSI Is Predicting Worlds Better Than We Think — But Not Well Enough to Call It a Rule

On the night of the MSI 2026 final, I sat in a rented apartment in Thu Dau Mot, Binh Duong, and reopened a spreadsheet that had followed me for three years. Two columns only: the MSI champion, and the Worlds champion of the same year. Six rows. Six matches. I closed the laptop and thought about something nobody in the newsroom wanted to hear: a six-for-six streak is not enough to call a rule. It only shows that we have been asking the wrong question. Numbers never lie — we simply have not asked them the right question yet.

I follow professional League of Legends from the angle of a data person. My job is not to predict who wins. My job is to check whether a model is being inflated by crowd emotion. In 2026 I was sent to Russia as an analytical reporter for the World Cup, and I did something that made my colleagues laugh in my face: I staked my entire career on a probability model named Croatia. Croatia's expected goals stood at 2.3; England's at 1.1. Croatia won 2-1 after extra time. Croatia was not a miracle — it was well-managed variance.

A year earlier, I had hand-coded the data of 182 V-League matches from video and found that Long An had the lowest PPDA in the league, 7.8, conceding just 0.7 goals per game through lightning counter-attacks. I wrote a piece that a veteran coach dismissed as soulless statistics, but the young assistant at Binh Duong FC invited me to build a pressing map for the team. The V-League is a mess, but every mess has its own rules. Since then I have kept one habit: before trusting a number, ask what question that number was born to answer.

The question the community has asked all season is a tidy one: what does MSI 2026 tell us about Worlds 2026? For years, the default answer from analysts was: not much. Winning MSI had never been a guarantee of winning Worlds. The mid-season event is played on a different patch, with a different schedule, and, most importantly, with a gap long enough for a team to lose its form.

In the last three years, the data has told the opposite story. JDG, BLG, Gen.G and T1 are four names that keep repeating across both arenas. MSI was expanded: more teams, more series, more best-of matches. MSI results now line up with Worlds results at a rate the old format never reached. That is why the community has started to believe MSI has become a reliable indicator.

I do not object to that optimism. I just want to know what that indicator is actually measuring.

First, take the number out of the story. Six for six sounds very strong, but the sample is only three years and four teams, all from the LPL and the LCK. No wildcard team appears on the list. No third region has broken in. In other words, the six-for-six streak does not prove that MSI predicts Worlds. It may only prove something more mundane: during those three years, the two strongest regions in the world remained the two strongest regions in the world. We think we understand the game — until the data opens our eyes.

I fell into exactly this trap once, in May and June 2026. When the pandemic paralysed competitions, I spent six weeks breaking down 252 Bundesliga matches played without crowds. Home win rates dropped from 43% to 29%; away teams ran about 6% more. The whole sports world cited that comparison table as scientific evidence of home advantage. Read closely, though, and the data only describes a systemic shock, not the nature of home advantage. Applause in an empty stadium records a truth nobody wants to hear. The six-for-six MSI streak is being read in exactly that way.

Based on my experience watching matches, the mechanism is the part the tables do not display directly.

The expanded MSI raises the information content of the tournament in three ways. First, the number of series increases, which lowers the probability that a team advances deep on luck. In a tournament with only a handful of best-of series, a favourable draw can carry a weak team to the semi-finals. In an expanded format, winning the title requires beating several strong teams in a row. Second, the double-elimination bracket gradually removes teams with thin rosters. Third, and this is the key point, the expanded MSI forces teams to prepare more draft structures in less time.

All three of those are skills that Worlds tests. When MSI measures the same skill set under the same kind of pressure, its results naturally align with Worlds more closely. This does not mean MSI predicts Worlds causally. It means the two tournaments have become structurally similar as tests.

One detail gets skipped: second and third place at MSI. The community only counts champions, but if MSI really is an indicator of depth, then the teams finishing just behind the winner are the group worth tracking at Worlds. Over the past three years, that group consistently went deep at Worlds, and in one year reached the final. An indicator you only read for the champion is an indicator used poorly.

And here I have to argue against myself. In data work, people forget one sentence: correlation is not causation. Six for six can be explained by a single variable nobody notices: roster stability. T1 has kept Faker in mid lane across multiple seasons; Gen.G is built around Chovy; BLG is tied to Knight; JDG once placed Ruler in the bot lane of its MSI-winning roster. Four teams, four cores kept nearly intact, four coaching staffs held together. If that is the real variable, then MSI is merely where that variable surfaces early, not the cause of success at Worlds. By the same logic, a team that wins MSI with a patchwork roster will struggle to repeat the feat.

The second blind spot concerns the patch. No public data shows how similar the tactical environment is between MSI and Worlds. We assume the two tournaments are played in a comparable environment, and that assumption has never been verified. One major patch landing between the two events could invalidate the entire six-for-six streak in a single season. If that happens, do not be surprised: that is missing data, not a wrong prediction. Many models collapse for that reason and no other.

Statistically, six observations is a sample far too small to name a rule. A fair coin tossed six times has about a 1.6% chance of landing on the same side six times. That figure is small, but it is small in a way that misleads people. Toss six times and see six heads, and you will believe the coin is rigged. Watch six tournaments and see six matches, and you will believe the same thing. Both cases are the same reasoning error.

I am not writing this to dismiss the value of MSI. I am writing to reset the question. The right question is not whether MSI predicts Worlds, but which kind of team MSI filters out. The preliminary answer: teams with roster depth, stable coaching systems, and the ability to adapt in a short window. That is a narrow set, and precisely because it is narrow, it easily overlaps with the group that wins Worlds.

What I will track at MSI 2026 is not the champion. I will track two things. One: who finishes second and third, and how far they go at Worlds. Two: whether any team outside the LPL and the LCK reaches the leading group. If a team from outside those two regions wins MSI, the six-for-six streak breaks in the most interesting way possible, and then we will know what this indicator actually measures.

And if MSI 2026 ends with another familiar name, remember that we still only have seven rows of data. Seven rows are not a rule. Seven rows are a hypothesis waiting to be disproved.