Why a well-formatted résumé with nothing in it is held down

Rewritten on 26 September 2026. The capped example this page was first written around, measured on 12 September, no longer meets a ceiling: since 25 September two or more achievement lines under a Projects heading count as experience evidence, and that document now scores 70 with no cap. The example below is one that meets a ceiling in today's checker, measured on 26 September. The ceilings themselves are unchanged.

Any scoring system that adds up formatting can be gamed by a document with formatting and nothing else. Clean headings, a readable contact line, one paragraph under a heading, and the sum looks respectable. The reader of that report then goes to an interview believing the hard part is done.

The ATS résumé checker closes that gap with four caps applied after the weighted sum, and we built a document to trip one.

The document we built

Mara Quill is invented: an operations coordinator whose résumé is a name, an email, a phone number, a city, and a three-sentence summary under one heading. It is well formed and readable, and it says who she is. It contains no role, no date, no degree and no skill list. Plenty of first drafts look exactly like this, because the summary is the part people write first.

What the report said

Parseability 92, completeness 30, content 2. The multiplication came to 37 and the published score was 25, and the line says why in words: limited to 25 because no experience or education was found. That clause is the whole diagnosis, and it points at the right half of the document: the page reads, and there is almost nothing on it to read.

The four caps

Under 25 readable words with no experience, education or projects, the score is held at 15. With no experience, no education and no skills found anywhere in the text, 25. With no experience and no education, 40. With no experience and a content dimension under 30, 55. They are applied in that order, after the sum, and only the first one that matches takes effect. Mara's page has 47 readable words and no skills section, so the second cap is the one that holds her.

The floor case

We also checked a four-line draft: a name, an email, a heading and one sentence. Parseability was held at 25 rather than the 100 its clean formatting would otherwise have earned, completeness reached 20 on the identity and the email, and content scored zero. The sum came to 14, under the floor rule's 15, so the line simply ends in the score.

The same person, uncapped

What lifts a ceiling is evidence, not formatting. Written with her roles under an Experience heading, dated, with achievement lines, a page like Mara's stops meeting any cap, and the score is the weighted sum again. On 12 September we showed this with a career written up as projects that was held at 40 and cleared the cap once the same work sat under an Experience heading; since 25 September the projects version clears it too. Either way the work did not change. The document did. Before and after are kept side by side under Previous checks, which is the subject of what a fix was worth.

Reading a capped report in the right order

Three things, in this order. First the arithmetic on the score card: when it ends in "limited to", a ceiling is in force and the weighted total is not what you are being judged on. Second the reason after it, which names what is missing. Third the checks under the score marked Review, each of which says what would earn its points back.

On Mara's report those three said: 37 weighted and 25 published, no experience or education found, and experience, education and skills all marked Review. That is one instruction rather than nine: put the career on the page. The counts are worth a glance too, because a capped score on a nine-word extraction is a file problem rather than a writing problem, which is the distinction drawn in what the checker read.

What a ceiling is not

It is not a judgement of your career, and it is not a prediction that an employer will reject you. It is a refusal by a measurement to report a number its evidence cannot support. The checks keep showing what each one earned while the cap is in place, so you can see both the work to do and the reason the number is not moving yet. The counts it made that decision from are printed as well, in what the checker read, and you can watch the whole sum assemble in why the checker prints its own arithmetic.