What is LIX and how is it calculated?
LIX is a readability index combining sentence length with the share of long words. The formula is words ÷ sentences + 100 × long words ÷ words. A long word has more than six letters. A higher score reflects longer sentences or more long words under these counting rules; it is not a universal measure of comprehension.
For example, “Elephants sleep. Cats play.” has four words, two sentences and one long word. Its LIX is 4 ÷ 2 + 100 × 1 ÷ 4 = 27.0. The display rounds to one decimal place; the calculation uses the full counts.
Exact counting rules
A word is an uninterrupted sequence of Unicode letters, combining marks and numbers, starting with a letter or number. Spaces and hyphens split words. An apostrophe inside a word does not, so “don't” and “aujourd'hui” are one word each. A long word contains more than six letters; number tokens are words but digits do not add to the long-word count. Accents are normalized so composed and decomposed Unicode letters count alike. Emoji and punctuation are not words. This is a practical token rule, not a grammatical analysis.
A period, exclamation mark, question mark or ellipsis ends a sentence. Consecutive ending marks make one boundary. A blank line also makes a boundary, while one line break does not. A final fragment without ending punctuation counts as a sentence; empty fragments do not. Periods inside abbreviations, initials, dates and decimals still split sentences. A heading without punctuation followed by a single line break may therefore join the next sentence. Use a blank line to separate it.
How to interpret a LIX score
Compare drafts of similar texts with the same counting rules, and inspect the counts alongside the score. There are different LIX bands in different educational contexts, so this calculator assigns no universal grade level. Topic, audience, structure and prior knowledge also affect comprehension. LIX arose in a Scandinavian language context. English vocabulary and German compounds can change the share of long words, so the same number need not mean the same reading difficulty across languages.
Input is limited to 20,000 JavaScript UTF-16 code units; an emoji may use two. Blank or wordless text has no score, avoiding division by zero. Review headings, lists, numbers, abbreviations and raw transcripts before interpreting a result. The text is calculated in the tab and is not put in a form, URL or tool log.