When a Korean title comes back to a chart it had left

Most of what looks like a return is a new season

How this is counted. One cell is one title in one market. A gap is a run of weeks in which the title is absent from that market's top 10, between two weeks in which it is present.

The first thing we saw was a title returning after 223weeks away, and the second thing we saw was that the obvious cases were second seasons. Netflix labels each chart row with a season, so we split the gaps on that label rather than guessing. It changes the answer.

⚠ The third row is not a small print. Netflix carries a season label on 29,791 of the 37,750 Korean chart rows between 2021-07-04 and 2026-07-26 — 78.9% — so 21.8% of gaps cannot be classified at all. We leave them in their own row rather than folding them into either answer.

The hypothesis we had, and had to drop

The obvious explanation for an old season charting again is that a new one arrived and pulled it back. We checked: for each genuine return, was another season of the same title on that chart within 4 weeks either side? It was in 16 of 310 cases — 5.2%. 94.8% of returns happen alone.

That is the finding, and it is not the one we expected. Old Korean seasons re-enter charts without a new season beside them. What brings them back is not in this data — see the last section — but it is not, in the main, a sequel.

Four weeks is our threshold, so here are the others

Four weeks is our threshold, not Netflix's. The table of other thresholds is published beside the result so the choice can be judged rather than trusted.

Read the third column down: it barely moves — 430 at two weeks, 412 at twelve. A gap between seasons is long by construction, so raising the threshold does not remove those cases. The fourth column falls by more than half over the same range. Whatever threshold you pick, the long gaps are mostly seasons.

How common is any of this

Returning is the exception. Most title-and-market cells are a single unbroken run, and the 54 titles that came back somewhere are a small part of the catalogue — which is why we are reporting a mechanism rather than a trend.

Where returns happen — and the gradient that dissolved

The obvious follow-up is whether comebacks cluster in the markets where Korean titles are already common. Counted per cell, they appear to: the rate roughly triples from the smallest Korean charts to the largest. That reading does not survive one control.

A market where Korean titles are common holds them on its chart for longer, and a longer run has more weeks in it — more opportunities for a gap to open. Divide by the number of adjacent week pairs rather than by the number of cells and the fourth column, which ranged from 1.4% to 4.8%, flattens to between 1.00% and 1.47%.

The last column is printed to two decimal places on purpose. Rounded to one, the first two bands both read 1.0% and the flatness would look like a rounding accident rather than a measurement — the same collapse we are warning about, one digit further along.

Returning is about as likely per opportunity everywhere. Had we published the per-cell column alone we would have sold a story about big markets reviving catalogues, and it would have been an artefact of run length. The collector now refuses to write if the per-opportunity figures spread by more than 1.5 points, so if that ever stops being true this page will stop building rather than quietly change its mind.

The longest gaps that were genuine returns

Alphabetical, not ranked. These are the ten longest gaps that survived the season test, shown in name order because this is not a league of comebacks.