The Daily Measure
Our methodology

Clear facts begin with a visible method.

We combine primary evidence with independent reporting, compare the claims source by source, and make uncertainty explicit. AI helps us process more perspectives consistently; it never becomes the source.

Evidence first

Every publishable claim must trace back to a source that a reader can inspect.

Balance, not false symmetry

We present the strongest factual case on each side without pretending weak evidence is equal.

Confidence, explained

A score describes how strongly the central facts are verified, not how certain the future is.

Our source hierarchy

No outlet is perfectly unbiased.

We therefore do not crown one publication as “the truth.” We favor sources with transparent standards, strong correction practices, direct access to events, and a record of separating reporting from opinion. For consequential claims, no single source is enough.

Story selection and factual verification are separate steps. Broad, sustained coverage across independent major outlets is a strong signal that a development belongs on the news agenda, especially when it fits a reader’s selected topics or regions. Coverage volume is not proof: the claims inside the story are checked separately before publication.

  1. Primary evidenceOfficial statistics, laws, court records, election authorities, company filings, research papers and direct statements.
  2. Independent wire servicesReuters, the Associated Press and Agence France-Presse form the usual factual backbone.
  3. Strong regional reportingLocal expertise is added so a global edition does not mistake one region’s lens for a universal view.
  4. Specialist reportingSubject experts and specialist publications add depth in science, climate, economics and technology.
  5. A genuinely different sourceWhen facts are disputed, we seek the strongest credible evidence that could change the conclusion.

Government, company and campaign sources can be essential primary evidence, but are never treated as independent corroboration of claims about themselves. Several outlets repeating the same wire story or press release also count as one underlying source, not several.

Reuters Trust Principles ↗AP News Values and Principles ↗AFP editorial charter ↗Ad Fontes methodology ↗

Verification confidence

A repeatable score, not an AI hunch.

Each story receives a claim-by-claim source record. Five parts are scored, added to 100 and rounded to the nearest five. The score is reviewed whenever material new evidence changes the central facts.

Evidence quality and directness30 points

Official records, raw data, filings, court documents, research papers and direct on-record evidence score highest.

Independent corroboration25 points

The central claim is confirmed by sources that did not simply copy one another.

Agreement on central facts20 points

Credible sources agree on what happened, while meaningful disagreements are made visible.

Event verification15 points

The time, place and participants are verified separately from the article’s publication time.

Freshness and completeness10 points

The evidence is current enough and sufficiently complete for the wording used.

85–100 High confidence70–84 Well supported50–69 DevelopingBelow 50 Normally not published

Urgent claims below 50 may appear only when their uncertainty is the news, and they must be plainly labeled unconfirmed.

Why use AI?

More comparison. Less emotional framing.

AI can compare many reports quickly, identify the facts they share, flag contradictions, translate material, and apply the same neutral structure to every story. That makes it useful for a wider and more consistent briefing than one person could produce each morning.

AI can also make mistakes, misread context or sound more certain than the evidence allows. It does not witness events and it is not a source. The published wording must remain traceable to evidence; high-risk stories involving war, elections, allegations or public safety require stricter review and corroboration.

The workflow

From the world to five stories.

1. Scan the news agenda. We compare major international outlets, wire services, relevant regional reporting and primary sources. Sustained coverage across independent outlets puts a story high on the candidate list; it does not make every claim true.

2. Weigh, rather than filter. Public consequence, international or regional significance, current media attention, the reader’s interests, available evidence and the balance of the full edition are considered together. A developing story does not need to be completely resolved to qualify.

3. Choose five without creating a bubble. Matching widely followed stories receive high priority. Preferences guide emphasis and order, while an exceptionally consequential world story can still appear outside them.

4. Verify claim by claim. Every central factual claim is checked against at least three reliable sources, preferably including primary evidence and two genuinely independent reports. Confirmed facts, attributed claims and unresolved details remain visibly separate.

5. Write facts before interpretation. The opening states what happened; plain language makes it readable; supported figures and context show scale. Loaded language, unsupported consequences and a prescribed “why it matters” conclusion are removed.

6. Show uncertainty and the record. Readers see what remains unsettled, the source count, confidence level and full source list. Material corrections should update the web edition and be disclosed in the next relevant briefing.

Clear facts. Honest balance. A wider world.

See the method in an edition