Sturgeon's Law
Learn it → Apply it → Save as Skill
Ninety percent of everything is mediocre. That isn't a complaint about any one field — it applies just as well to science fiction, software, résumés, or research papers. The real question isn't whether a field has a lot of weak output; it almost certainly does. It's whether you're comparing its best against someone else's best, or judging the whole field by the majority nobody bothered to filter out.
Learn it. Apply it. Save as Skill.
The Dr. Wallace’s Favorite Mental Models is designed to turn useful ways of thinking into reusable tools. Each model gives you a fast visual overview, deeper interactive learning, practical application exercises and an AI-ready Markdown skill you can save for future use.
Instead of merely reading about mental models, you can gradually build your own AI skills library: a collection of portable reasoning skills that can help ChatGPT, Claude, Grok and other AI assistants apply the mental models you value.
In any field with a large number of independent attempts, most of the output will be mediocre by default — not because the field is uniquely bad, but because effort and skill vary enormously and most attempts land somewhere between average and weak. Dismissing or defending an entire field by its worst (or best) visible examples is the wrong move. The fix: expect the mediocre majority, build a real filter, and compare the best against the best.
Quality doesn't spread evenly — it clusters
The weak majority isn't evenly spread, and neither is the strong minority. A raw, unsorted pool of submissions, listings, or candidates is dominated by noise. One deliberate filter pass removes most of it. What survives in the top decile is usually concentrated among a handful of sources — not spread evenly across everyone who tried.
Use it before you judge, invest in, or dismiss a whole field.
The model is most useful right before you decide whether an unfamiliar genre, platform, or category is worth your time, before triaging a large pool of applicants or submissions, or before deciding how to build a curation or screening process.
Expect it → Compare top-to-top → Build a filter → Find the cluster.
Sturgeon's Law is often remembered as a cynical one-liner. Applied well, it's a four-part discipline for evaluating any large, open pool.
Assume most output in any open pool — submissions, applicants, listings, papers — will be weak by default. That's normal variance in effort and skill, not evidence the field itself is broken.
When judging whether a field, genre, or option set has value, compare its best examples against the best of something else — never its worst against someone else's best.
Don't wade through the majority hoping to stumble onto the good fraction. Build or borrow a deliberate filter — screening criteria, editorial review, peer review, curation.
The valuable minority is rarely spread evenly across every source. It tends to cluster among a small number of creators, vendors, or candidates — locate the cluster instead of averaging the whole pool.
Six things worth understanding before you apply it.
Expand each to see the exact wording, why it was written, and the well-documented ways it gets applied wrong.
The Exact Statement
What Sturgeon actually wrote
The line first appeared in print in Sturgeon's book-review column in the March 1958 issue of Venture Science Fiction: "Sure, ninety percent of science fiction is crud. That's because ninety percent of everything is crud." The word most people quote today — "crap" instead of "crud" — is a later popular substitution, not Sturgeon's original wording. The Oxford English Dictionary traces the underlying idea to a talk Sturgeon gave at New York University around 1951–52, and the line spread through science-fiction fandom after he repeated it at the 1953 World Science Fiction Convention.
Why He Said It
Defending a genre judged by its worst examples
Sturgeon was responding to critics who dismissed science fiction as a genre by pointing to its worst, pulpiest examples. His point wasn't that science fiction was secretly all great — it was that singling out a field's worst output and judging the entire field by it is a bad method, because applying that same test to literary fiction, journalism, or any other field would produce the same result.
A Technical Wrinkle
Is this really "Sturgeon's Law"?
Sturgeon himself reportedly reserved the name "Sturgeon's Law" for a separate, different quip: "Nothing is always absolutely so." The 90%-crud observation was originally called "Sturgeon's Revelation." In practice, almost every modern reference — including this page — uses "Sturgeon's Law" for the 90% statement, since that's now the overwhelming popular convention. Worth knowing if an older source draws the distinction differently.
Why It Tends to Hold
A production-economics observation, not a formal law
Nobody has measured this precisely, and "90%" is a rough stand-in for "the large majority," not a statistic. The pattern shows up because, in any field with low barriers to entry, a huge number of people attempt something, effort and skill vary enormously, and most attempts land somewhere between mediocre and bad — ordinary variance, not a special flaw in any one field.
Where It Breaks Down
High barriers to entry change the ratio
Fields with real barriers to entry — peer-reviewed journals, licensed professions, juried competitions, capital-intensive industries — filter out much of the weak output before the public ever sees it, so the surviving pool looks far better than 90% weak. Applying the 90% figure to an already-filtered sample, like one issue of a top journal or a curated best-of list, misreads the law.
The Modern Operationalization
Filtering at scale
Platforms with huge volumes of user-submitted content — app stores, self-publishing platforms, video platforms, and now AI-generated content — build ranking and recommendation systems explicitly because raw, unfiltered browsing is dominated by low-quality material. Editor's picks, curated charts, and ranking algorithms are Sturgeon's Law operationalized as infrastructure.
Two hundred applications land for one open role.
This is an illustrative example designed to teach the method.
The same mediocre-majority, concentrated-minority pattern shows up well outside publishing.
Sturgeon was defending science fiction, but the underlying pattern applies anywhere a large number of independent attempts meet real variance in effort and skill.
Millions of listings, a tiny visible fraction
The overwhelming majority of apps in major app stores are abandoned, broken, or barely functional. Curated charts and editor's picks exist because browsing the raw catalog is mostly noise, not because the stores lack good software.
Most papers are rarely cited
A large share of published academic papers receive few or no citations after release. Systematic reviews and meta-analyses exist specifically to filter the literature for quality and relevance, rather than treat every published paper as equally worth reading.
Output concentrated in a small creator base
Across video, self-publishing, and music platforms, audience attention concentrates heavily among a small percentage of creators, while a vast majority of uploads get negligible engagement — the good fraction clustering, not spreading evenly.
Judged by outliers, not the median company
Venture investors routinely see the large majority of pitches and portfolio companies underperform or fail outright. The asset class is still judged by its rare, concentrated outlier successes — not by averaging across every company that was pitched.
Fair use of the law, or misuse?
Each scenario below applies — or misapplies — Sturgeon's Law. Pick which one it is, then check yourself.
1. A critic dismisses a specific, well-researched complaint about a weak sequel from a usually excellent studio by saying, "That's just Sturgeon's Law — 90% of everything is crap," without engaging with the actual complaint.
2. A reader assumes most unreviewed self-published novels will be mediocre, so builds a filter — checking sample chapters and verified reviews — before deciding whether the self-publishing scene has anything worth reading.
3. An editor argues that one issue of a highly selective, award-winning journal must "also be 90% crap," since that's supposedly just how everything works, without accounting for the journal's selective review process.
4. A founder isn't discouraged about entering a crowded market, reasoning that most existing competitors in any crowded field tend to be mediocre, and a genuinely well-built product can still stand out.
A 1958 book-review column that outlived the column, the magazine, and almost the author's own term for it.
Expand for the origin of the model, the man who wrote it, and a wrinkle most popular sources skip.
Who was Theodore Sturgeon?
Theodore Sturgeon (1918–1985), born Edward Hamilton Waldo on Staten Island, New York, was an American science fiction and fantasy author. His best-known novel, More Than Human (1953), grew steadily in reputation after publication. He also wrote two aired episodes of the original Star Trek series — "Shore Leave" and "Amok Time" (1966–67), the episode that introduced the Vulcan hand salute and the line "Live long and prosper." He was inducted into the Science Fiction Hall of Fame in 2000.
Where exactly did the 90% line come from?
It first appeared in Sturgeon's book-review column in the March 1958 issue of Venture Science Fiction, where he wrote: "Sure, ninety percent of science fiction is crud. That's because ninety percent of everything is crud." The Oxford English Dictionary traces the underlying idea to a talk Sturgeon gave at New York University around 1951–52, and the phrase spread through science-fiction fandom after the 1953 World Science Fiction Convention.
Is the 90% figure meant literally?
No. It's a rough stand-in for "the large majority," not a measured statistic — Sturgeon never claimed to have measured anything. "Crud" (or "crap") is also subjective: people disagree about what counts, and quality within both the valuable fraction and the weak majority varies enormously. Treat the number as a reminder to expect a lot of mediocrity by default, not as a precise ratio.
The line is simple; applying it responsibly is not.
Citing it as a conversation-ending defense
Using the law to dismiss a specific, substantive criticism without showing the criticized example is actually unrepresentative of the field.
Applying it to an already-filtered sample
Assuming a top journal, juried competition, or curated best-of list must also be 90% weak, when its selection process already removed most of the weak material.
Assuming the good fraction is evenly spread
Treating every source, creator, or candidate as equally likely to be in the valuable fraction, when it's usually concentrated among a few.
Treating "90%" as a measurement
Reading the number as a precise statistic rather than a rough way of saying "most."
Using it to excuse your own mediocre work
"Everyone's output is mostly weak, so mine is fine too" skips the actual point, which is to filter harder — not to lower your own bar.
Ignoring how barriers to entry change the ratio
Applying the same 90% expectation to a licensed, juried, or capital-intensive field that it was built to describe in an open, low-barrier one.
Some situations need a different default.
Use rigorous head-to-head comparison when...
you're evaluating an already-filtered, high-barrier sample — a shortlist, a juried competition, a peer-reviewed journal — where assuming 90% of it is weak is simply wrong.
Use it as a search strategy, not a verdict, when...
deciding whether to engage with an unfamiliar field at all — build a filter and look at its best examples, rather than dismissing the whole field from the outside.
Use honest self-assessment when...
evaluating your own work — the law describes other people's output in aggregate; it isn't a license to skip quality control on what you personally ship.
One model rarely does the whole job.
Build a Sturgeon's Law filtering prompt.
This browser tool structures your thinking. It does not send your information anywhere or perform AI analysis.
Your AI-ready prompt
Turn this mental model into a reusable AI skill.
Save as Skill converts the model into portable Markdown instructions that teach an AI when and how to apply Sturgeon's Law, what questions to ask, what evidence to seek, what mistakes to avoid and how to structure its analysis.
Build your own Mental Models Skills Library
Save the skills you find useful and organize them into your own reusable library. The Markdown is intentionally model-neutral, so it can be used wherever an AI system accepts uploaded files, project knowledge, custom instructions or reusable text context. Platform capabilities differ, so this is a portable skill file rather than a claim of one-click installation.
For readers who want the nuance.
Does Sturgeon's Law mean quality is hopeless to find?
No. It means the valuable fraction exists but is a minority you have to filter for deliberately — it's a case for building better filters, not for giving up on a field altogether.
Why do some sources say "Sturgeon's Law" isn't really about the 90% line at all?
Because, strictly, it isn't. Sturgeon reportedly used "Sturgeon's Law" for a separate aphorism, "Nothing is always absolutely so," and called the 90%-crud observation "Sturgeon's Revelation." Popular usage has since merged the two under the more famous name, which is the convention nearly everyone, including this page, now follows.
Does this apply to AI-generated content?
Increasingly, yes, as a contemporary extension rather than part of Sturgeon's original 1958 argument. As the volume of AI-generated text, images, and code grows, the same pattern applies: most of it will be mediocre by default, and filtering and curation matter more, not less, as raw volume increases.