Three in five Korean film titles share their name with a foreign work. For series it is one in four.

Our Southeast Asia panel holds 397 Korean titles. A title is in it because its English name matches a Korean work in Wikidata. That rule can fail in exactly one way: another country made something with the same name.

So we asked, for every title in the panel, which countries produced a film or series carrying exactly that label.

Three in five of our film titles cannot be settled by name. For series it is fewer than one in four. Anyone using our film numbers is standing on much softer ground than anyone using our series numbers, and until now we had not said so, because we had not measured it.

The reason is length, mostly

Sort the same 397 titles by how many words are in the name and the pattern is not subtle.

A one-word title is more than six times as likely to collide as a four-word one. Carter, Karma, Hunt, Believer, Ballerina — each of those is a Korean work, and each is also the name of something else, often several things.

Korean films get short names. The median film title in the panel is two words and 40% of them are a single word; the median series title is three words and only 14.8% are one word. Series arrive with names like Extraordinary Attorney Woo and When Life Gives You Tangerines, which nothing else in the world is called.

But length is not the whole gap

Hold the length fixed and the films are still worse.

At every length, a Korean film’s title is roughly 25 points more likely to be shared than a Korean series’ title of the same length. Short film names are not just short, they are drawn from a smaller and more common pool — one-word abstractions, single names, ordinary nouns.

This means our film panel is the part of our data most likely to contain something that is not Korean, and it is the part we can least easily check. It is also the smaller side: 160 titles against 237, and 449 chart-weeks against 1,810.

A second ruler we had not used

Wikidata tells us a collision is possible. It cannot tell us which work actually charted. For that we had one instrument sitting unused: Netflix’s own South Korea chart.

A Korean title normally also plays in Korea. Of the 149 shared titles, 76 have never appeared on Netflix’s Korean top ten in five years. Twenty of those charted in exactly one country in the world, and fourteen of the twenty are films.

That list contains the clearest miss we have found since the title-matching correction:

• Impetigore charted for five weeks in Indonesia and nowhere else on earth. It is an Indonesian horror film. Wikidata has a Korean work of the same name, so our rule took it.
• The Lord Musang King charted for one week, in Malaysia. It is Malaysian.
• Wildflower charted for thirteen weeks, in the Philippines, in no other country, ever.

One of those three is no longer in the panel. Between this piece first publishing and this revision we removed eight titles, Wildflower among them, for a reason stronger than the queue: our own attribution query returned no Korean work at all for those exact names. Impetigore and The Lord Musang King stayed, because Wikidata does list a Korean work of each name. Both are still in the queue and both, we think, are the Indonesian and Malaysian films.

Why we have not simply deleted the eighty-four

Because the same test also flags titles that are obviously Korean.

Vagabond charted eight weeks in Vietnam and nowhere else, including not in Korea. It is a Korean drama. Smugglers and Keys to the Heart are Korean films; neither ever reached Korea’s own top ten. A Korean title can miss the Korean chart for reasons that say nothing about where it was made: it was released before July 2021 and had its run before this data begins, it was a cinema film Netflix never streamed domestically, or it simply lost its week at home to something bigger.

Absence from the Korean chart is a reason to look, not a verdict. The measurement produces a queue: 20 titles that charted in one country only, 19 more concentrated above half in a single country, 37 that spread widely but never touched Korea. We will work down it and every removal will appear on our corrections page with the evidence that moved it.

What this changes about reading us

Two things, and we would rather you took them from us than found them yourself.

Our series figures are the solid half. 23.6% name ambiguity, and the biggest series in the panel are Netflix originals whose names are unique.

Our film figures carry real risk of contamination, concentrated in one-word titles that charted briefly in a single country. Those are also the smallest entries — the median flagged title stayed two weeks — so the risk sits in the tail rather than the top. It is still there.

The one thing we can say without hedging is the size of what we do not know, which is what this piece is. A panel that reported 397 titles and said nothing else would be a cleaner-looking number and a worse one.