# Ellipsis usage in writing: how it is scored and how to change it > Ellipsis usage records how often a writer uses ellipses, on a four-step scale: never, rare, occasional or frequent. Canonical: https://scriptgrain.com/attributes/ellipsis-usage # Ellipsis usage in writing *By Jack Stovell · published 2026-09-24* Ellipsis usage records how often a writer uses ellipses, on a four-step scale: never, rare, occasional or frequent. A ScriptGrain profile stores the extraction model's label; when a draft is scored, code counts its ellipses per 1,000 words and converts the rate to a label. At least half of the 299 pieces in ScriptGrain's reference corpus have none. ## Analyse your ellipsis usage free Your free voice profile measures ellipsis usage among all 45 attributes, from a sample of your own writing. No card needed. [Try the six-signal check first](https://scriptgrain.com/#sg-measure) ## At a glance - API field: **ellipsis_usage** - How it is measured: **Measured in code** - Group: **Punctuation and format** (ScriptGrain voice profile, 45 attributes in 8 groups) - Scale: **frequent, occasional, rare or never** (ScriptGrain attribute definitions) - Voice match: **Scored, weight 1** (ScriptGrain voice-match engine (published method)) - Median in published writing: **0 ellipses per 1,000 words (the rate behind the label)** (ScriptGrain reference corpus: 299 public pieces, built 20 September 2026) ## What ellipsis usage measures Ellipsis usage is simply how often ellipses show up in your writing, expressed on a four-step scale: frequent, occasional, rare, or never. The app's own definition is blunt: "How often ellipses (…) appear: frequent, occasional, rare, or never." It sits in the [punctuation and format](https://scriptgrain.com/attributes/punctuation-and-format) group with comma density, [exclamation rate](https://scriptgrain.com/attributes/exclamation-rate) and [question marks in body](https://scriptgrain.com/attributes/question-marks-in-body). When your voice profile gets created, an extraction model (Claude Sonnet 5, running in two passes) reads your writing samples and forms a judgement about your ellipsis habit, along with 44 other attributes. That judgement is snapped to one of the four labels at save. When a *draft* gets scored, code takes over: it counts every "..." and "…" per 1,000 words and converts the rate to a label with fixed thresholds. The free habits report shows the same rate against the corpus. Two mechanisms, same four labels. ## How ellipsis usage is scored Ellipsis usage does move your voice match score, at a weight of 1. When a draft gets checked against your profile, code counts the draft's ellipses per 1,000 words, converts the rate to a label (never, rare, occasional, frequent) and compares it with your stored label on an ordered scale. Land on the same step and you score 1. One step apart (say, occasional versus rare) scores 0.5. Two or more steps apart drops you to 0.1. A stored label the scale cannot read scores 0.15. A writer whose profile reads "never", scored against a draft full of ellipses, gets 0.1 on this feature. Full details on how the weighting and calibration work sit in the [voice match](https://scriptgrain.com/glossary/voice-match) glossary entry and the [voice measurement framework](https://scriptgrain.com/reference/voice-measurement-framework). ## How to analyse ellipsis usage in your own writing You don't need software to get a rough read on your own habit. Here's how: 1. Pull a writing sample of at least a few hundred words, ideally something typical of your usual output rather than one unusual piece. 2. Count every instance of "..." and every instance of "…" as one ellipsis each. 3. Divide the total count by the number of words in the sample, then multiply by 1,000 to get your rate per 1,000 words. 4. Compare that rate against the thresholds: 0.01 or less per 1,000 words is "never", under 0.8 is "rare", under 2.5 is "occasional", 2.5 or more is "frequent". 5. Repeat across two or three samples if you can. One piece can mislead you. That's the manual version. ScriptGrain's free voice profile reads this attribute from a writing sample among all 45 attributes. ## Typical ellipsis usage in published writing Most published writing barely touches the ellipsis. ScriptGrain's reference corpus is 299 public pieces from 47 sources (835,655 words), built 2026-09-20, each measured in code and then discarded. Its ellipsis rate per 1,000 words, the rate behind the label, sits at: p10 is 0, p25 is 0, median is 0, p75 is 0.98, and p90 is 2.33. Read that plainly. At least half of the pieces in the corpus contain no ellipsis at all. The 75th percentile, 0.98 per 1,000 words, already reads as "occasional" (0.8 or more), and even the 90th percentile, 2.33, falls short of the 2.5 that makes "frequent", so fewer than one piece in ten reaches it. If your own writing sits at zero, you're not an outlier. You're the median. ## How to make AI use ellipses like you If your writing genuinely leans on the ellipsis and you want AI output to match, there are a few concrete moves. 1. Allow a trailing ellipsis specifically where you leave a thought open, rather than scattering them at random. 2. Decide which form you actually use, three dots or the single "…" character, and stay consistent. Mixing both looks careless. 3. Show the AI real examples pulled from your own writing so it can copy the pattern rather than guess at it. When a ScriptGrain profile generates a draft, the ellipsis usage label goes inside the voice profile JSON block of the system prompt, under "Treat every attribute as a hard constraint, not a suggestion." There's no special ellipsis rule beyond that. The same profile reaches ChatGPT, Claude and other tools through the API and the MCP server. ## How to make AI stop using ellipses If AI output keeps sprinkling ellipses into your writing and you want it gone, the fix is direct. 1. Ban ellipses outright, except where they're marking words genuinely omitted from a quotation. 2. Finish trailing thoughts as full sentences instead of trailing off. 3. Replace any dramatic pause with a full stop. It reads cleaner, and it forces the sentence to actually land. The same mechanism works in this direction: when your samples read as "never", that label goes into the voice profile JSON as a hard constraint with every ScriptGrain draft. ## Questions ### How often should you use an ellipsis? Rarely, if the reference corpus is any guide. At least half of the 299 public pieces in ScriptGrain's corpus use no ellipsis. The 75th percentile is 0.98 per 1,000 words, which already counts as "occasional", and fewer than one piece in ten reaches "frequent" (2.5 or more per 1,000 words). Unless your voice depends on trailing thoughts, treat the ellipsis as a rare tool, not a habit. ### What does an ellipsis mean in writing? It signals an omission or a trailing thought: words left out of a quotation, or a sentence that deliberately doesn't finish. In ScriptGrain's framework, ellipsis usage just measures how often those marks appear (frequent, occasional, rare, or never), not what any single instance means stylistically. The label describes a habit across your writing, not the function of one particular ellipsis. ### How is ellipsis usage scored? It carries a weight of 1 in the voice match score. Code counts your draft's ellipses per 1,000 words, converts the rate to a label, then compares it with your profile's stored label on an ordered scale: same step scores 1, one step apart 0.5, two or more steps apart 0.1, and a stored label the scale cannot read 0.15. See the voice measurement framework for the full method. ### How do I check my ellipsis usage with an ellipsis checker? Count every "..." and "…" in a sample, divide by word count, multiply by 1,000. That gives your rate per 1,000 words. Apply the thresholds: 0.01 or under is "never", under 0.8 is "rare", under 2.5 is "occasional", 2.5 or more is "frequent". That's a manual ellipsis checker. ScriptGrain's free voice profile gives the extraction model's own reading of the habit, as one of its 45 attributes. ### How do I make ChatGPT use ellipses the way I do? Tell it explicitly which form you use (three dots or "…"), keep that form consistent, and show real examples of where you trail off versus where you don't. A ScriptGrain profile stores your ellipsis label: its own drafts send it in the voice profile JSON as a hard constraint, and ChatGPT can read the same profile through the API or MCP server without you writing the instruction by hand each time. ## Related - [Punctuation and format attributes](https://scriptgrain.com/attributes/punctuation-and-format) - [Comma density](https://scriptgrain.com/attributes/comma-density) - [Exclamation rate](https://scriptgrain.com/attributes/exclamation-rate) - [Question marks in body](https://scriptgrain.com/attributes/question-marks-in-body) - [Punctuation and formatting as a fingerprint: dashes, commas, list habits](https://scriptgrain.com/reference/punctuation-and-format) - [micro-habits](https://scriptgrain.com/micro-habits) ## Sources - [How ScriptGrain scores voice match](https://scriptgrain.com/reference/voice-measurement-framework) - [The attributes and their corpus percentiles](https://scriptgrain.com/reference/writing-voice-attributes)