Okinawa has elected a new governor: Genta Koja, a newcomer, defeated the incumbent Denny Tamaki, who was seeking a third term. I counted 3,645 comments: 90 to 10. An expert note printed above the thread called this Okinawa's first social-media governor election, and 80 comments answered it directly. First place took 35,565 presses. Comments that regretted the result came to 59, one tenth of one percent of the vote, and the highest-placed of them sat at 143rd

What happened

At 19:59 on Sunday 13 September 2026, RBC Ryukyu Broadcasting published this story on Yahoo! News Japan. Okinawa voted for a governor that day. Genta Koja, 42, a newcomer and until recently deputy mayor of Naha, is certain to have won. Voting closed at 8pm. Counting had not started, but RBC called the race on the strength of exit polling at the booths, campaign surveys from the early part of the race, early-voting patterns and its own reporting.

The story gives the winner's record. Born in Naha in 1983. A private high school there, then the University of Tokyo, then the Ministry of Internal Affairs and Communications as a career official. Back in Okinawa in 2022, and in July of that year, at 38, a run for the upper house of the national parliament: 270,000-odd votes, close to the sitting member, but a loss. Deputy mayor of Naha from that December, resigning this February. This was a first run for governor, with the backing of six parties.

The pledges, too. A "new 5K economy" — tourism, health, the environment, the sea and entrepreneurship — meant to more than double income per head in the prefecture from its level three years ago, to five million yen, along with a bigger childcare budget. On the bases: accept the plan to move the US Marine Corps air station at Futenma to Henoko, and press for the return of other US facilities south of Kadena.

Some background for readers outside Japan. Okinawa is a prefecture of islands at Japan's southern end, and a great many US military facilities sit there. Futenma is a US airfield surrounded by housing, long described as dangerous. The plan to move it to a place called Henoko, in the city of Nago, is the dispute that has set the prefecture against the national government for years. The phrase "All Okinawa" does not appear in the story, but 161 comments use it: it is the name for the cross-party grouping built around opposition to that move. "Zero-uchi" is another thread word — a broadcaster calling a race the moment the polls close, before a single ballot is counted. That is what happened here.

One more thing, and it matters for how this page reads. Yahoo! News sometimes prints commentary from named specialists directly above the reader comments. This story carried two. One, from an economist, said that raising incomes is the urgent task and that what is wanted is practical management of the economy and of the bases rather than an argument about ideology. The other became the largest stone in the thread.

My own reading

Here is what I wrote when I read the story.

[…] people on the mainland have mostly cast this side as the villain.

Loud, and never delivers […] and when people like this explain why they lost, they blame social media.

Twelve years of it.

There are two sets of brackets above but three passages missing, and I should say what they were. The first is a whole opening sentence, asserting as fact a tie between an elected official and a foreign political party. The second and third sit inside the brackets in the middle, asserting as fact things about that same person's handling of money and about their character. This site does not abuse a named individual in the body of an article, and that holds for my own words as much as anyone's. So all three are out, and I am recording here that they are out.

That leaves three threads. The mainland has cast this side as the villain. Loud, and never delivers. And: when they explain a defeat, they blame social media.

The third one landed. What I did not expect was where it landed — not in the losing camp, but in the paragraph printed above the comments.

90 to 10 — 3,645 comments, 11 hours after the polls closed

I counted at 7:23 on the morning of 14 September, 11 hours after voting closed. There were 3,645 top-level comments; the story showed 5,009 comments in all. The presses came to 178,560 of I relate, 4,375 of that makes sense and 19,281 of hmm. Ninety to ten.

It is a thread that agrees with itself, but the shape of the vote is unlike any I have counted. The top 10 comments took 83 percent of it, the top 20 took 89 percent, the first-placed comment alone took 35,565 presses, 20 percent, the top two took 35 percent and the top five took 64 percent. Two thirds of this thread's mood is carried by five comments. Below them it goes quiet: of the 3,645, some 866, 24 percent, drew no I relate at all, and 1,117, 31 percent, carry one press or none. There were 926 replies in the thread, and 268 of them hang off first place.

So read what follows as two numbers at once, as always: how many people wrote it, and how many people pressed it. This time those two numbers part company at exactly the point that matters.

Two things to keep in mind while reading the counts.

The first is the arithmetic of years. The story says the incumbent was seeking a third term, which makes it two terms, eight years. The thread, though, contains 12 comments that say twelve years, four of which also name the previous governor: counting from 2014, three terms of the same grouping in office. Seventeen comments name that previous governor in all, while 92 contain the word for eight years, so the thread mostly counts in eights.

The second is when this thread was written. The race was called the moment the polls closed, before a single ballot had been counted. So most of these comments were written by people who had no idea how wide the margin was. A few say as much: let us wait for the count to finish and see what the gap turns out to be. Only 24 comments mention turnout at all, with 376 presses between them. What this thread is working from is not the size of the win. It is the bare fact that the answer arrived at the same instant the polls shut.

First place: we have finally pushed them out

First place, 35,565 presses, 20 percent of the vote. It also drew the most hmm in the thread, 3,171, and the most replies, 268. The most-pressed comment was also the most pushed back against.

The two incidents it names — a collision involving a works vehicle at a place called Awa, and the capsizing of a protest boat off Henoko — appear nowhere in the story. The thread brought them in. Comments raising them numbered 174, drawing 45,997 presses, 26 percent of the vote. I cannot establish the facts of either from this one story, so I record only that the thread wrote about them. Some comments named the person who died. This site does not publish the names of private individuals.

The comment I shared on X

Second place, 27,321 presses, 15 percent of the vote. Last time, in Leave it at home, the comment shared on X placed seventh and took 3 percent, so this is a good deal higher. Four short sentences. Where first place lists incidents and gives reasons, this one gives no reason at all, and it still reached nearly four fifths of first place's vote. Short, flat statements do well in this thread. It drew 1,852 hmm and 115 replies, second in the thread on both.

"Okinawa's first social-media governor election" — the note, and the thread's answer

Here is the centre of it. The second expert note above the comments read, in substance: this can be called Okinawa's first social-media governor election. More posts than in last year's Miyagi governor race came from outside the prefecture, and a great many defamatory posts and hoaxes are said to have spread. The established media inside the prefecture led on fact-checking, but against the wave of posts, did that work really reach voters here? Was the result swayed, to the end, by that information war?

The thread reacted to those few lines more than to the story. Comments touching on social media, the internet, hoaxes, defamation or an information war numbered 283, 7.8 percent of the thread, and drew 18,740 presses, 10.5 percent of the vote. Comments mentioning the note itself numbered 160, with 2,391 presses.

Of those, the ones that answered it head-on — saying that blaming the defeat on social media insults the voters — numbered 80, with 2,080 presses, 1.2 percent of the vote. The highest of them is 13th.

Thirteenth place, 1,437 presses. Where it sits is the interesting part. Places one to twelve are all some version of we won, this is good, they are out. Then, at thirteen, comes a rebuttal of a single paragraph printed above the comments. And notice what the comment does not do: it does not deny that hoaxes existed. If there were hoaxes, it says, criticise them — but once you make them the explanation for the result, you have turned a decision into a symptom.

The comment in 34th place puts the same thing in fewer words: how many people actually voted because of abuse online? Making that the reason for defeat is shifting the blame outward, and it feels like underestimating the voters.

Read all 80 and the vocabulary barely changes. Insult. Shifting blame. Treating us as fools. Look at the eight years first. In counts that is 2.2 percent of the thread and in votes 1.2 percent, which is not much. But nowhere else in this thread do so many people write so nearly the same sentence.

I should add that many of the 80 carry abuse aimed at the person who wrote the note. Those I count and do not quote. The two quoted here argue without it.

There is a comparison worth making here. Two expert notes were printed above these comments. The first, from an economist, said that raising incomes is the urgent task, that what is wanted is practical management of the economy and of the bases rather than a fight about ideology, and that the problem has to be solved without splitting the prefecture. Nobody answered that one. Of the 160 comments that mention the expert notes at all, not a single one in the whole 3,645 names the economist or their institute. The other note is named by 127. The same readers, in the same thread, let one note pass in silence and wrote 80 replies to the other. The difference is not the politics of the two notes. It is that one of them contains a sentence that can be read as: you lost because of what you are like.

The other side of the argument is in the thread too. Comments saying the spread of hoaxes is the frightening part, and that fact-checking could not keep up, are there: one at 162nd with 18 presses of I relate against 13 hmm, another at 466th with 5 against 14. They were pushed back.

A comment from someone who voted

Fifth place, 14,641 presses. Among the top comments it is one of the very few that says outright: I voted. Comments that identify themselves as living in Okinawa numbered 37, with 14,889 presses between them — and almost all of that is this one comment.

I stopped at one phrase in the middle: I refused to be carried along by the hoaxes. This was written before the expert note appeared. So here is someone insisting, unprompted, that they were not misled. Something during the campaign had already made that worth saying.

Counted, not quoted

Comments mentioning China or Taiwan numbered 365 and drew 38,068 presses, 21 percent of the vote, and plenty of them sit high in the thread. But most of them attack the outgoing governor by name. Abuse of a named individual does not go into the body of an article here, so I give the count and do not quote them.

Comments mentioning foreign residents numbered 14, with 22 presses. Comments mentioning the Battle of Okinawa numbered 8, with 40 presses. Both counted, neither quoted.

Comments containing words such as activist, left-wing or Communist Party numbered 443, with 41,040 presses, 23 percent. Comments naming a local newspaper or broadcaster and criticising its coverage numbered 100, with 17,146 presses, 10 percent. Comments containing the word for congratulations numbered 539, with 54,985 presses, 31 percent — the commonest word in the thread.

The minority: 59 comments, 197 presses, none above 143rd

Picking out by hand the comments that regretted the result, opposed the new direction, or argued that Futenma will not come back down this road, I found 59. Between them they drew 197 presses of I relate, one tenth of one percent of the vote, and 506 hmm — more pushback than support. The highest-placed of the 59 sits at 143rd. Not one reached the top 140.

636th place: four presses of I relate against 30 of hmm. First place in the same thread took 35,565.

At 204th, a comment objects that the moment the race was called the thread filled with one side only, that not every voter in Okinawa voted that way, that it is rude to someone who worked for eight years, and that what happens in Okinawa is for the people of Okinawa to decide — 13 presses of I relate against 18 hmm. At 143rd: I respect the result, but I expect an administration that cannot say no to the government, and will the water testing around the bases really continue? Twenty-two presses against 55 hmm.

Several of the 59 carry abuse of their own. Those I count and do not quote.

One comment belongs beside the minority without being part of it. The most pushed-back comment in the whole thread is not about winning or losing. At 215th, 12 presses of I relate against 82 of hmm, it says that calling a race before the polls close makes a mockery of the voter's right and saps the will of anyone who has not yet voted. On that one point the winning side and the losing side pushed back together — because this thread was delighted the answer came early.

Checking my own reading

I had three lines.

First: people on the mainland have mostly cast this side as the villain. This thread cannot settle that. It was written after the result and leans 90 to 10 towards the winners, so it is no evidence of how the mainland saw things beforehand. Unknown — and unknown is what I write when a thing cannot be counted.

Second: loud, and never delivers. Comments containing the word for eight years numbered 92. Comments on the economy or daily life, 260, 8 percent of the vote. Comments noting that this prefecture has the lowest income per head in the country, 34, 3 percent. The comment in 14th place writes that the question in the end became whether the incumbent had delivered economic growth, and that the answer was no — then closes by congratulating the winner and thanking the loser for their work. Right.

Third: when they explain a defeat, they blame social media. This is the one the thread bore out, and it did so in a shape I did not expect. The first place that explanation appeared was not the losing camp. It was the paragraph printed above the comments, by a named specialist, within hours of the polls closing. Two hundred and eighty-three comments reacted; 80 answered it directly. Several comments also say they watched the concession press conference and heard the same explanation there. I did not watch it, so I record only that the thread wrote so.

Two right, one unknown.

One last thing. What drew the most presses in this thread was the joy of winning. But what was repeated most often, in almost identical words, was a protest: by your account of this, we did not think about it at all. Underneath 90 to 10 there was a second argument running the whole way down.