Direct answer: Of 41 cached article pages, 28 have an H1 and 13 do not; 32 carry a visible date. Among the 27 pages with both an H1 and a date, 20 place the date after the title (median gap 409 characters, 7 within 200), while 7 place it before (E14 and E23, median 797). Another 5 pages have a date but no H1, and 8 have neither. A date near the title is not the default.
Study 30 counted whether dates exist; study 33 located them in the page; this one asks the sharper question: how far is the date from the H1, and is it before or after the title?
Method: approximate character offsets
Pages come from the 41 successful crawls in study 29. Scripts, styles, comments and tags are removed to obtain visible text; the first H1's start and end are mapped into that text, as is the first date. A date before the H1 is before, after it is after, and the gap is counted in characters.
| Field | Definition | Boundary |
|---|---|---|
| H1 | Visible text of the first <h1> | Recorded as no when absent; quality not judged |
| First date | First 20xx numeric date in visible text | List, copyright and body dates all count |
| Direction | Date before or after the H1 | Before: date appears first |
| Gap | H1 start to date (before) or H1 end to date (after) | Approximate, affected by whitespace folding |
Result: 20 after the title, 7 before
| Item | Pages | Note |
|---|---|---|
| With an H1 | 28 | 27 have a date, 1 does not |
| Without an H1 | 13 | 5 have a date, 8 do not |
| Date after the H1 | 20 | Median 409 chars; 7 ≤200, 11 ≤500 |
| Date before the H1 | 7 | Four E14, three E23; median 797 chars |
| Gap after title | Pages | Note |
|---|---|---|
| ≤200 characters | 7 | Date immediately below the title |
| 201–500 characters | 4 | After the lead |
| >500 characters | 9 | Mid-body or footer |
Seven pages put the date before the title
Four E14 pages and three E23 pages show their first date before the H1, with gaps from 351 to 4,377 characters. A common cause is an article template that puts “published” metadata above the title, or a list/breadcrumb appearing before the body heading. The date is still visible, but the order is “date → title”, not “title → date”.
These pages are a reminder that “near the title” cannot be replaced by “early in the page”. Being early does not mean being after the heading; direction has to be recorded separately.
Thirteen pages have no H1
Thirteen pages lack an <h1>: five still carry a date, eight have neither an H1 nor a date. An H1 is a content-structure signal and an entry point for search and AI systems; missing it does not remove dates, but it makes “date relative to the title” undefined.
Acceptance should record two things separately: structure (is there an H1) and date (does one exist, before or after the title). The eight pages missing both deserve their own line.
What this means for acceptance
- Check the H1 before the date. Thirteen of 41 have no H1, so date-to-title distance cannot exist there.
- Record direction separately. Twenty after the title and seven before mean different things.
- “Near the title” is not the default. Even after the H1, the median gap is 409 characters.
- Character gaps are not pixels. Font size, line height, images and accordions change the real distance.
- List the eight pages missing both. Do not merge structural absence with date absence.
Limits
- Approximate offsets: whitespace folding and tag stripping introduce small errors.
- First H1 only: pages with multiple H1s use the first; hidden H1s still count.
- No JavaScript: script-injected titles and dates are invisible.
- No semantics: list, copyright and body dates all count as the first date.
- 41-page pilot: no industry claim.
Data and reproduction
Anonymised data: equipment-title-date-distance-2026.csv (41 rows: sample_id, URL, H1 flag, date flag, H1 length, date count, first date, H1 start/end, direction, gap). The repository keeps docs/research/2026-11-title-date-distance.csv.
Parser: scripts/title_date_probe.py; unit tests: scripts/test_title_date_probe.py (15 tests, no network). Run:
python -m unittest scripts.test_title_date_probe
Related: the date position pilot (median first date at character 402), the visible dates pilot (32/41 with dates), the visible date checker, and the privacy and analytics pilot (signals on 25 pages).
Sources and definitions
- Page cache: 41 successful pages from study 29; raw HTML kept off-repository.
- Position: visible-text character offsets; no rendering, no JavaScript.
- Scan: 2026-11-02, zero network requests.