How scoring works

Significance is an estimated importance rating from 0 to 10. It is not a factual accuracy or source credibility rating. Automatic AI processing is disabled. No numeric significance scores or measured publisher-coverage counts were available in the live audit on 23 October 2026. Not scored means no assessment is published; it does not promise a scheduled assessment. Score unavailable means source text, permissions or supported impact evidence are insufficient, or a result awaits review. Headline-only stories omit the score badge.

What the estimate considers

• People meaningfully affected
• Severity of consequences
• Geographic reach
• Expected duration
• Change from the existing situation

Consider these factors together, using only information the permitted excerpt supports. Unknown reach or duration stays unknown. Severe consequences for a small group can still matter. Popularity, sensational wording, publisher identity and article length do not automatically raise a score.

A score is an editorial estimate of the reported event, not an objective measurement or a rating of factual accuracy or source credibility. Unknown impact is unknown; it does not establish low importance or contained consequences. Return null when source evidence cannot support an estimate. Supporting quotations are checked against the supplied text to aid review; a matching quotation does not prove that a generated claim is faithful or that the original reporting is accurate. Editorially reviewed summaries are checked against the supplied excerpt, not independently verified news.

Announced decisions and proposals may matter before implementation when the source documents their scope and potential consequences. Preserve conditions and uncertainty; supported potential impact is not a completed outcome.

Scoring anchors

• 0–2: Source evidence establishes very limited practical consequences. Example: A routine change explicitly reported to have little practical effect.
• 3–4: Documented meaningful but modest consequences, with limits supported by evidence. Example: A service disruption explicitly reported to affect one site for a few hours.
• 5–6: Documented substantial consequences or a lasting change for affected people. Example: An announced service change with explicitly documented substantial effects if implemented.
• 7–8: Documented major consequences with severe or sustained effects. Example: A decision with source-supported severe, sustained effects on essential services.
• 9–10: Documented exceptional, severe and durable consequences or fundamental systemic change. Example: A source-supported emergency with exceptional, widespread and lasting consequences.

These examples illustrate the scale; they are not scored real stories. Rubric: significance-v3. Short excerpts can omit vital context, and model judgments can be inconsistent or biased. Insufficient evidence leaves the summary or score unavailable. AI summaries can contain errors: check the original reporting. Demo scores remain illustrative.

Sorting the feed

• Significance: highest estimated significance first; equal scores use original publication date, then stable ID. Unscored stories come last.
• Coverage: highest explicitly measured number of distinct publishers first; unknown coverage comes last. Multiple articles from one publisher count once.
• Latest: most recent original publication date first; unknown dates come last.

Coverage indicates reporting breadth, not accuracy or importance. We do not infer event matches or coverage counts. Ties are resolved by a stable story ID.