Return to the evidence as it was captured.

Atlas currently preserves one verified reading. It does not invent a time series where none has been captured.

Rank and discussion are different measures.

12 ranked stories0 missing values

Ask, Show, and Jobs over time

Coverage boundary

How long a story stays

Unit: hours on the pageRanks 1–30 of the top list

Median hours on the front page by the best rank a story reached
Peak rank	Stories	Median hours

How this is measured

What a second attempt does

Unit: share reaching the thresholdAssociation, not effect

Why this is not an effect

The shape of a discussion

Unit: replies per opening comment, and depth against depthRule: each thread read against the same captured baseline

Discussions read against the captured baseline, with what was captured of each
Discussion	Shape	Opening comments	Replies each	Deepened	Comments captured

What the thread answered

Posted about, or talked about

Each subject’s share of titles beside its share of comments
Subject	Share of titles	Share of comments	People who wrote them

How this is measured

Duration × score change

Loading 36 records

Waiting for the complete historical observation file.

Cohort: current Ask/Show/Jobs rowsObserved: retained adjacent intervals ≤20mSpan / segments: first to last; restart after gaps or observed absenceAbsence: not evidence of HN removal

Every selected story, separating counted observation from its full first-to-last span
List	Rank	Story	Score	Comments	First seen	Observed	Full span	Segments	Snapshots	Best rank	Score Δ	Comment Δ

No second verified reading

Cross-day comparison stays unavailable until another verified snapshot is intentionally added.