With 86% of the words, you might understand only 38% of an anime's lines

The analysis argues that vocabulary “coverage” can greatly overstate how much anime dialogue a learner can follow. Coverage counts every known word occurrence, so frequent particles and grammar words inflate the score; understanding a line, by contrast, requires knowing every word in it. In the first 3,000 lines of Frieren, the 500 most frequent words cover 85.9% of spoken words but occur together in only 38.4% of lines. The gap reflects lines averaging about seven words: even a high chance of recognizing each word leaves many lines with at least one unfamiliar term. The measure is demanding, but it makes clear why a strong coverage figure may not feel like easy viewing.

The same pattern appears across eight series. With 500 show-specific words, the share of lines understood ranges from 32.7% for Jujutsu Kaisen to 55.9% for Serial Experiments Lain; reaching half the lines takes 401 words for Lain, 730 for Frieren, and 980 for Jujutsu Kaisen. Many missed lines are close: in Frieren, 56.4% lack just one known word, so expanding from 500 to 1,000 words raises understood lines by 22.1 percentage points, despite a 6.5-point coverage gain. The figures come from subtitle lines tokenized and ranked by frequency, and define understanding as knowing every word. They are optimistic: they ignore grammar, context, pronunciation, multiple word meanings, and whether learners actually know the selected vocabulary.