Self-Edit Like an Editor: The Reverse-Outline Trick for Catching Padding

The reverse outline editing technique is simple: after you finish a draft, write one short line for what each paragraph actually does — and any paragraph you can’t summarize is padding. Where a normal outline is a plan you write before drafting, a reverse outline is an x-ray you take afterward, and it exposes the repetition, filler, and gaps that fluent sentences hide.

Editors do a version of this instinctively. They read for structure, not just prose, because they know a piece can be beautifully written sentence by sentence and still be a mess — saying the same thing three times, wandering off topic, padding a thin idea with generic sentences. The reverse outline is how you catch all of that in your own work, where you’re too close to see it.

Key takeaways

• A reverse outline is made after drafting: one line summarizing each paragraph’s real job.
• Any paragraph you can’t summarize in a line is probably padding — cut or merge it.
• It moves you from sentence level (where good prose hides bad structure) to structure level.
• AI drafts pad heavily because models generate fluent filler endlessly; the outline exposes it.
• Do it after a complete draft but before final line-editing — fix structure first, then polish.

Why you can’t see your own padding

When you re-read a draft, you read it the way you wrote it — one sentence at a time, pulled forward by the prose. If the sentences are fluent, you glide right over the structural problems. Two paragraphs that make the same point don’t announce themselves; you read the first, nod, read the second, nod again, and never notice you just nodded at the same idea twice. A paragraph that’s pure filler still contains real words and correct grammar, so at sentence level it passes.

Padding is a structure problem, and you can’t catch a structure problem from inside the sentences. You have to get above them. That’s the entire trick: the reverse outline forces you up to the altitude where the shape of the argument is visible all at once.

How to use the reverse outline editing technique

The mechanics take about ten minutes on a normal article.

1. Number your paragraphs. Or just go top to bottom.
2. For each one, write a single short line — five to ten words — capturing its one main point. Not what it’s *about*; what it *does*. “Argues detectors output probabilities, not proof.” “Gives the coffee-shop example.”
3. Put the lines together in a list. Now you’re looking at your actual structure, stripped of the prose that was disguising it.
4. Read the list as an argument. Does it flow? Does each line advance the piece? Are any two lines nearly identical?

The magic is in step 2. When a paragraph resists a one-line summary — because it has no single point, or restates the paragraph above, or just re-announces the heading — you’ve found a problem. A paragraph that can’t be summarized usually isn’t doing work.

What the list reveals

Once your draft is a list of one-liners, three problems jump out that were invisible in prose:

• Repetition. Two lines that say the same thing. One of those paragraphs goes.
• Padding. A line that’s vague (“discusses the importance of the topic”) or that just restates the section header. That paragraph is filler; cut it or give it a real job.
• Gaps. A jump in the logic — line 4 assumes something line 3 never established. Now you know where to add.

You can’t unsee these once they’re on a list. That’s why the technique works: it converts a structure problem into a visible, obvious one.

A worked example

Say you reverse-outline a draft and get this:

> 1. AI writing often sounds robotic. > 2. This is a problem for students and professionals. > 3. Robotic writing is a common issue for many people. > 4. Here’s how to fix it: vary your sentences. > 5. Varying sentences is important for good writing. > 6. Step two: cut the jargon.

Reading the *prose*, this might feel fine — every sentence is competent. Reading the *outline*, the problems are glaring. Lines 1, 2, and 3 all say “robotic writing is bad and common” — that’s one point stretched over three paragraphs, pure padding. Line 5 just restates line 4’s importance without adding anything. The real content is lines 4 and 6, the actual steps. A tightened outline:

> 1. AI writing sounds robotic — a real problem for students and pros. > 2. Fix one: vary your sentence length. > 3. Fix two: cut the jargon.

Three paragraphs instead of six, and nothing of value lost. The reverse outline turned a padded draft into a lean one by making the redundancy impossible to ignore. This kind of structural tightening pairs naturally with a 15-point self-editing checklist for the sentence-level pass that comes after.

Why this matters for sounding human

Padding is one of the clearest signatures of unedited AI text, and there’s a mechanical reason. A language model can generate fluent, on-topic, grammatical sentences forever without adding a single new idea — restating the thesis, listing generic benefits, explaining why the topic “is important.” It fills space effortlessly because filling space is easier than saying something. The result is the thin, repetitive, low-information texture that makes unedited AI content read as thin.

Cutting padding raises the information density of every paragraph — the ratio of real content to word count — and high information density is one of the most human things about good writing. It’s also what Google’s helpful-content guidance rewards: pages that actually inform rather than restate. A human editor cutting a draft is ruthless about this; the reverse outline lets you be ruthless about your own.

Be honest about the detection angle: thinner, padded text does tend to read as more machine-like, and denser writing as more human, but detectors output probabilities, not verdicts, and they misjudge honest work — OpenAI retired its own classifier in 2023 for weak accuracy. Cut padding because a lean piece respects the reader’s time, not to move a number.

The workflow

1. Finish the draft first. Reverse-outlining a half-draft tells you nothing.
2. Summarize each paragraph in one line. What it does, not what it’s about.
3. Read the list of lines as your real structure. Ignore the prose for now.
4. Kill repetition, cut padding, fill gaps. Merge duplicate points, delete filler paragraphs, add where the logic jumps.
5. Then line-edit what survives. Polish sentences only in the paragraphs that earned their place.

Closing

You can’t fix a structure you can’t see, and fluent prose is very good at hiding a bad structure from its own author. The reverse outline pulls you up above the sentences to where the padding, repetition, and gaps are obvious. Ten minutes of summarizing each paragraph in a line will do more for a bloated draft than an hour of polishing sentences that shouldn’t survive the cut.

Once you’ve tightened the structure and want to check whether the surviving prose still reads as machine-made, run the draft through PaperBleach and see where the flags land. More self-editing guides are in the PaperBleach blog.

Frequently asked questions

What is a reverse outline?

A reverse outline is an outline you make after writing, not before. You go through your finished draft and write one short line summarizing what each paragraph actually does — its single main point. The result is a bird’s-eye map of your real structure, as opposed to the structure you intended. It exposes padding, repetition, and gaps that are invisible when you’re reading the prose sentence by sentence.

How does a reverse outline catch padding?

When you try to summarize a paragraph in one line and can’t — because it has no single point, or repeats the paragraph above, or just restates the heading — you’ve found padding. A paragraph that resists summary usually isn’t doing work. The reverse outline forces every paragraph to justify its existence in a few words, and the ones that can’t earn their place get cut or merged. It makes filler visible in a way close reading doesn’t.

Does padding make writing look AI-generated?

Yes, notably. AI drafts are prone to padding because models can generate fluent, on-topic sentences indefinitely without adding new information — restating the point, listing generic benefits, filling space. That produces the thin, repetitive, low-information texture that reads as machine-made and that detectors partly respond to. Cutting padding raises the information density of every paragraph, which is one of the most human things you can do to a draft.

How is a reverse outline different from just re-reading my draft?

Re-reading keeps you at sentence level, where good prose can disguise bad structure — you get pulled along by fluent writing and miss that two paragraphs make the same point. A reverse outline forces you up to structure level, where you see the whole shape at once. Problems that are invisible up close (repetition, missing logic, a section in the wrong order) become obvious when the whole argument fits on one screen as a list.

When should I do a reverse outline?

After a complete draft exists but before final line-editing. There’s no point polishing sentences in a paragraph you’re about to cut, so fix the structure first: reverse-outline, spot the padding and gaps, reorder and delete, and then do your sentence-level editing on what survives. For short pieces you can do it in your head; for anything long or important, write the outline out where you can see the whole thing.