With 86% of the words, you might understand only 38% of an anime's lines
Summary
The article analyzes anime subtitle word coverage vs lines understood, using a 500-word top list to measure how much of dialogue a viewer can follow. It explains the methodology (Python, Sudachi tokenizer), presents results across eight shows, and discusses factors like polysemy and line length that separate coverage from actual comprehension. It includes charts, tables, and a public code recipe to reproduce the analysis.