Six Out of Six: The MSI–Worlds Scoreboard and What Three Years of Data Has Not Said Yet

On the night of the MSI 2026 final, the esports cafe in the Seomyeon district of Busan was full. I sat in the last row, next to a group of university students, and as usual I was not watching the match the way other people watch a match. I was watching the secondary screen.

On it, the analyst desk displayed a two-column graphic. The left column read MSI champion, the right column read Worlds champion, and each row joined a name on the left with a name on the right. Six rows. All six matched. The host said something to the effect that history was leaning heavily to one side.

At the bottom of the screen there was a very small footnote. I paused the frame, zoomed in, and still could not read it. The type was about six points, projected through three yellow bulbs in a Busan cafe. That footnote, the line explaining the criteria by which those six rows had been counted, was the only thing I took home that night.

It took me eleven days to finish counting again. And I did not count six.

The context of a scoreboard

MSI was born in 2026 in Tallahassee, Florida, originally imagined as a gathering of the spring champions of each region. For the first seven editions, in 2026, 2026, 2026, 2026, 2026, 2026 and 2026, with the 2026 edition cancelled by the pandemic, the tournament kept a compact format: a round-robin group, a knockout bracket, one team lifting the trophy. From 2026, Riot Games expanded the format, added teams, introduced a play-in stage and lengthened the main bracket. At the same time, MSI results began to carry weight in Worlds seeding. That is when the story of MSI predicting Worlds appeared.

I have a personal tie to this tournament. MSI 2026 was held in Busan, at the BEXCO exhibition centre, twenty minutes by metro from where I live. It was the first major international esports event in my city after two years of disruption, and I sat in that arena for four straight days, taking notes on every pick and ban. I learned something there that I still repeat to young editors: between the real arena and the virtual one, only the name differs, not the heart. The BEXCO stands were half empty that year, but the noise still reached the ceiling.

At BEXCO in 2026 I sat in row eleven, section B, and noticed something: when a big teamfight broke out, the shouting did not come from the stands but from the row of seats behind the technical area, where several dozen staff and substitute players were standing. An empty arena does not erase the cheering, it only moves it into our memory. I bring up that detail because it relates directly to how we remember MSI editions: what gets recorded in the data is often not what actually happened inside the building.

What I read, and what is circulating in the community, goes roughly like this: winning MSI usually does not mean winning Worlds, but after the MSI format was expanded, the correlation between the two tournaments became noticeably stronger; the six most recent rows of data all match, and all of them come from the LPL and the LCK, through JDG, BLG, Gen.G and T1. The author himself warns that three years of results are not enough to call this a rule. I agree with that warning. But I want to go one step further, into the place the scoreboard has never been willing to talk about.

The core: definition comes before statistics

When I paused the frame at the forty-second mark of that clip and counted by hand, I got four names, not six. Those four names were JDG, BLG, Gen.G and T1. To get six rows, whoever built the graphic had to count by a rule different from mine. Perhaps they counted the MSI finalists and not only the champions. Perhaps they counted the team that won the main bracket and the team that came through the lower bracket. Perhaps they merged two stages of the same MSI, something the expanded format allows if you treat each stage as a separate title.

There is no way for me to know for certain, because that footnote was unreadable.

I called three people. A Korean commentator who has worked at each of the last four MSI editions, a Chinese analyst specialising in the LPL, and a technical staffer who once worked on the graphics production for an international event. All three said the same thing: they did not know exactly which rule that graphic used, because such rules are decided by the production crew and are usually written late, while the graphic is being assembled, by someone who needs a round number on screen. None of them considered it misconduct. They considered it ordinary television work.

The problem is not that someone deliberately misled the audience. The problem is that, since 2026, the phrase MSI champion is no longer a single concept. There is a play-in stage winner. There is a lower bracket winner. There is a finalist. There is the team that wins the final. There is the team awarded seeding points. Each definition produces a different set, and each of those sets is then compared against another list, the list of Worlds champions, which can also be counted in more than one way. A statistic without a definition is not evidence; it is a poster. And a poster, however beautiful, does not stand up to a real match in October.

That is why I keep the habit of recording unusual details at every tournament I follow. Not to catch anyone out, but because those details tend to point straight at the place where a claim becomes thin. A six-point footnote is such a detail.

The base rate across ten editions

Now set the scoreboard aside and look at the whole history.

MSI has been held ten times, from 2026 to 2026, excluding the cancelled 2026 edition. In those ten editions, exactly once did the MSI champion also win Worlds in the same year: SK Telecom T1, in 2026. In every other case, the MSI champion watched a different team lift the Summoner's Cup in October. If you take that as the baseline, and a serious writer on numbers must take it as the baseline, then the probability that an MSI champion goes on to win Worlds is roughly one in ten.

So what changed in the last three years? There are two possibilities, and I cannot yet separate them with the available data.

The first: the expanded format really has made MSI a closer rehearsal for Worlds. More teams, more knockout matches, more days on stage. All of that forces a team to expose its real playbook instead of hiding it.

The second: three years is only three years. Three years is not a trend; three years is three coin flips, and fan memory only keeps the flips that landed face up. Against a baseline of one in ten, a run of three matching editions is not so strange that it demands a theory built around it.

The difference between these two possibilities is not academic. If it is the first, next year's scoreboard will be right again, and teams should treat MSI as a genuine indicator. If it is the second, next year's scoreboard will be wrong, and those who placed their faith in it will lose something.

Why two regions fill the whole scoreboard

I spent two evenings doing a simple count: between 2026 and 2026, what share of the titles at the two biggest tournaments belonged to the LPL and the LCK?

A very high share. And this is the point I want to make.

If those two regions win roughly nine out of every ten international titles, then taking an MSI champion and comparing it with the Worlds champions list will automatically produce a match for purely statistical reasons, not because MSI has prophetic power. You do not need any mysterious mechanism. You only need two regions so deep in talent that any team in their top tier sits in the contender pool of every event they attend. Names like Chovy, Faker, Knight, Ruler and Bin remain the guarantee of both leagues' drawing power, and also the guarantee that the two tournaments always share the same pool of contenders.

In other words, the overlap between MSI and Worlds is largely the overlap of the LPL and the LCK with themselves. The scoreboard does not measure the relationship between the two tournaments. It measures the concentration of talent in two countries.

But I do not want to push this argument too far. There is a subtle distinction I found in the very material I read: MSI is good at identifying contenders, but not good at identifying champions. Those are different jobs. Identifying contenders means that after MSI you can narrow the list of possible Worlds winners to five or six names. Identifying the champion means naming exactly one. Across ten MSI editions, narrowing the list has succeeded fairly often. Naming the one has almost never succeeded.

And that leads to a harder conclusion: if MSI only predicts the teams everyone had already predicted, then it has not predicted anything.

The real mechanism: seeding, brackets and a few weeks money cannot buy

There is a mechanism behind this correlation, and it is not at all mysterious. It is administrative.

Since the format was expanded, MSI results have been used to determine seeding and to allocate Worlds slots. Put plainly: a team that goes deep at MSI does not only earn prestige. It earns an easier bracket in October.

And there is something else, which I consider more important than seeding. A team that goes deep at MSI gets two to three weeks of matches on a real stage, against international opponents, with a full coaching staff and the highest-quality analytical data of the entire year. No summer scrim block reproduces that. You can buy players, buy coaches, buy facilities. You cannot buy the feeling of an international knockout match in game five.

MSI does not predict Worlds. MSI grants slots, grants brackets, and grants a roster a few weeks that no amount of money can purchase. Once you rewrite the correlation as a causal relationship like that, it becomes entirely ordinary, and it loses the glamour of a prophecy.

The cost sits there too. May to October is five months. A team that goes deep at MSI has fewer rest days, more travel, more interviews, and solves the spring tactical puzzle one step ahead of its rivals. In esports, where stamina shows up not in collisions but in reaction speed and practice hours, that gap is not small.

The biggest hole: the patch in between

And this is the part that bothers me most when I reread all the material I gathered.

Not one of those documents mentions the game version.

Nobody says which patch MSI 2026 was played on and which patch Worlds 2026 was played on. Nobody says how many champion changes, map changes, objective bounty changes or lane-swap rule changes occurred between the two events. Nobody says whether the update before Worlds erased the advantage MSI handed to its champion.

For a tournament where every result depends on the game version, failing to mention the version is a gap that collapses the whole argument. If you want to know how October will go, the first question is not who won MSI, but which playstyle the MSI champion excels at, and whether the September update still rewards that playstyle.

I followed a team in Busan in 2026 that won all four group games with a tactic built around a mobile mid laner. Three months later, an update significantly changed the strength of a few items, and that team went out in the first knockout round. Nobody on the roster got worse. The game changed the question.

People watch the MSI standings all through May; the thing that actually decides October is written in the update notes, in smaller type than the footnote on that scoreboard.

Second and third place: the most interesting spot on the board

In all the analysis I read, there was one line I consider the most valuable, and it sat at the edge: that finishing second or third at MSI can be as interesting as winning it.

I believe that, and I believe it for a very specific reason. In the expanded format, the runner-up and the third-placed team are usually the teams that play the most matches of the tournament. They come through the lower bracket, they win repeatedly under must-win pressure, and they do it against opponents who have already shown their hands. A champion often only needs to execute one plan over three weeks. A third-placed team often has to execute five different plans in the same window.

Every rough gem once lay still under the mud, waiting only for a patient enough gaze. The team that loses the final is such a gem. The standings record only the champion, but the history of a season is often written by the team that left the stage last without a trophy in hand. If this season one of the MSI runner-up or third-placed teams goes deep at Worlds, next year's scoreboard will need an extra column.

What I will track for 2026

I set myself three signals, and I wrote them down so that I cannot fool myself with hindsight later.

First, a team outside the LPL and the LCK winning MSI. If that happens while an LPL or LCK team still wins Worlds, the scoreboard loses half its rows.

Second, the performance of the MSI runner-up and third-placed teams. If they go deep, the hypothesis that MSI measures regional depth is strengthened. If they collapse early, the hypothesis that MSI merely measures poor concealment gets stronger too.

Third, and most important, the gap between the patch played at MSI and the patch played at Worlds. I want to read an analysis that states clearly: which version this year's event is played on, what the pre-Worlds update changes, and which playstyle those changes reward. Without that section, every scoreboard is decoration.

And a fourth signal, a small one, just for me: whether next year the broadcast will print the footnote large enough for viewers to read.

The counterintuitive angle

There is a harder version of this story, and I think it is partly true.

The MSI predicts Worlds narrative exists partly because the esports media needs a thread running from May to November. Between the two tournaments lie five empty months. No qualifier is compelling enough to fill that gap on a global scale. So the scoreboard appears: it gives people a story to follow all summer.

And this scoreboard has a property that makes it almost impossible to falsify in the short term. If MSI and Worlds match again next year, it is confirmed. If they do not match, people explain that the game version changed, and that explanation can, technically, always be true. A hypothesis that cannot be wrong cannot be used to predict either.

What the cameras do not capture is usually the thing most worth filming. In this case, what the cameras did not capture was the footnote. The eleven days I spent were not meant to prove the scoreboard wrong. They were meant to find out what it was counting, and I still do not have a complete answer.

What remains

If MSI only predicts the teams all of us had already predicted before the tournament began, then its real value lies elsewhere: it is the only test in the year that measures how fast an entire region reads the game version, rather than measuring a single roster. I will not write an ending for November. I do not write endings, I only go looking for the roads nobody has told yet. And if the scoreboard is right again this year, I will still be the first to go looking for its footnote.