Verify the Claims

Every factual claim should be traceable to one or more source records. The source record should show exactly what the evidence supports, whether the page is using it for a verified fact, analysis, scenario or limitation.

Mapping keeps broad war narratives from drifting away from evidence. It also lets readers challenge a sentence without disputing the whole page, because the claim, source and caveat are inspectable together.

The publication date tells readers when a source entered the public record. For war evidence, that date matters because casualty counts, reconstruction estimates, battlefield conditions, sanctions and nuclear-safety updates can change quickly.

A dated assessment should not be stretched beyond its own window. For example, a reconstruction figure tied to damage through 31 December 2025 should not be presented as the cost of later damage unless a newer source supports it.

The accessed date records when this site last checked the source. It is not the same as the source publication date, and it should not be used to make an old source look new.

Access dates help readers understand link reliability and review cadence. If an organisation revises a page, moves a document or changes a data table, the site can show when the earlier claim was last inspected.

Source type tells readers whether a record is a primary document, government release, international organisation report, research publication, journalism or NGO material. The category describes provenance; it is not a shortcut for automatic trust.

Different source types answer different questions. A treaty text can support a legal framework, an IAEA update can support a nuclear-safety status claim, and research can support modelled economic mechanisms if its assumptions are shown.

Confidence labels describe how strongly a source-backed claim can be used on the site. They consider source quality, specificity, currentness, corroboration and whether the wording stays inside what the evidence actually proves.

High confidence is not certainty about every hidden fact, and low confidence is not a claim that something is false. The label is an editorial signal about how carefully the claim should be reused.

Limitations explain what a source does not prove, where its data window ends and which assumptions shape the claim. They stop a true sentence from becoming misleading when copied without context.

This is especially important for verified minimums, modelled scenarios and reconstruction estimates. A verified casualty figure can understate total harm, while a damage assessment can be authoritative for one date and incomplete for later attacks.

Archived references preserve a view of a source at or near the time it was checked. They help readers audit a claim if the live page later changes, moves or becomes unavailable.

An archive does not replace the live source and does not make a weak source stronger. It is a durability tool for transparency, especially for public records, press releases and datasets that may be revised.

Correction history records what changed, when it changed and why. It should cover numerical corrections, source replacements, wording changes that alter meaning and updates to translated or downloadable material.

Corrections protect reader trust by making changes visible instead of silently overwriting the record. The goal is not to punish mistakes; it is to keep the evidence chain readable after publication.