What Is Burstiness?
Burstiness describes the degree of variation in sentence length and structure within a piece of writing. High burstiness means the text alternates between short and long sentences unpredictably — a twenty-word sentence followed by a four-word one. Low burstiness means sentences cluster around similar lengths with regular rhythm. Human writing, measured across large samples, tends toward higher burstiness; machine-generated text clusters at uniform lengths because models resolve each next-sentence decision toward the statistically average choice.
Detectors usually quantify it as the coefficient of variation of sentence lengths — the standard deviation divided by the mean — so five sentences measuring 13, 15, 14, 16, and 12 words score close to zero, while a thirty-one-word winding clause followed by a two-word verdict ("It failed.") scores far higher. Structure counts as well: a paragraph that opens with a claim, digresses once, and lands on a fragment varies more than one built entirely from subject-verb-object units of identical shape.
That makes burstiness one of the two pillars of AI-text detection (with perplexity) and, more usefully, one of the clearest signals you can edit for yourself. If your paragraph reads monotonously aloud, mechanically splitting or merging sentences raises burstiness directly — it's the single highest-leverage edit for making drafting feel less machine-flavored, which is why every serious rewriting pass targets sentence-length variety first.
Two misconceptions cause trouble. First, burstiness is not randomness for its own sake — a deliberately clipped instructional manual or a court ruling can be legitimately uniform, which is one more reason detectors misfire on formulaic human prose. Second, varying a single sentence accomplishes nothing, because the measurement reads the distribution across the whole passage rather than any one sentence. Writers revising AI drafts get outsized value here, since generated text defaults to medium-length declaratives: read the paragraph aloud, mark where your own voice would speed up or stop short, then cut or extend accordingly.
Example usage
"Her drafts always need a burstiness pass — every paragraph is five tidy fourteen-word sentences until she breaks them up." An editor uses the term this way during line edits to name a specific revision move: split the longest compound sentence, merge two short ones, and let one line end abruptly for emphasis.
Related Terms
See it in practice
Apply this concept hands-on with our free ai humanizer tool — no signup required.