Sat Aug 15 — Removing a single title left one of our medians exactly where it was and moved the other by 89% of itself. The interquartile range rated the two almost identically. We ran neither before publishing.
Mon Aug 10 — Two works can carry the same English title, and a chart row does not say which one it is. Thirteen titles in our Korean panel turned out to be foreign works. The test that found them is not the name — it is where they charted.
Sun Aug 09 — Korean titles hold 7.9% of the world's Netflix top 10 places — 39,612 of 498,480. Of those, 82.3% carry the label we use to exclude same-name foreign works. 16.4% carry no label at all, and we keep them because we cannot check them.
Sat Aug 08 — We tagged every changed figure with the kind of mistake that produced it. One kind accounts for eight of the thirty-three. Four kinds still have no test, and this piece names them rather than reporting the coverage.
Fri Aug 07 — Our Korean Netflix catalogue was 294 titles and 27.7 billion hours. It is 236 and 23.7 billion. Here is the mistake, how it spread, why counting alone would never have caught it, and what we changed so it cannot happen the same way twice.
We hold 95 tags in all. 33 of them are on a single piece, and we do not publish a page for those — the tag still shows on the article, it just does not become an address of its own.