Methodology

The daily Global Risk level summarizes the alerts CHRONOS published on your local calendar day. It is the median urgency level of that day's published alerts, rounded up to the next whole level. For example, a day with nine alerts at urgency L3, L3, L3, L4, L4, L4, L5, L5, L5 has a median of L4.

The day runs from midnight to midnight in your browser's timezone, including shortened or lengthened daylight-saving days. Only public alerts with a valid urgency level are counted; alerts published that day and later resolved still count, while retracted alerts are excluded. If an alert's urgency is revised, the current level is used. Page filters, categories and search never change the number.

Global Risk describes only the selected set of published alerts. It is not a measurement or prediction of all global conditions, and a single alert never sets the level on its own. A historical Daily Report shows the level for the Los Angeles day that report covers, which can differ from today's number.

Intelligence should be falsifiable.

CHRONOS ALERTS does not collect only the evidence that supports an interpretation. Each alert also attempts to document:

• negative evidence that weakens the thesis
• resilience and shock absorbers that could contain the impact
• confirmation signals that would strengthen it
• invalidation signals that would weaken it
• known uncertainty
• the specific evidence that would prove the interpretation wrong

Alerts aim to state what would strengthen or weaken the assessment. Required fields do not guarantee that every condition is complete or sufficiently specific; inspect the evidence and uncertainty for yourself.

CHRONOS ALERTS works to verify material information using credible sources and, when practical, cross-checks consequential claims against multiple independent or authoritative sources. Primary and official evidence is prioritized when available.

However, CHRONOS ALERTS can make mistakes. Sources may be incomplete, outdated, incorrectly reported, misunderstood, later corrected, disputed, or wrong.

Automated analysis may misunderstand a source, miss important context, incorrectly connect events, misclassify information, overstate or understate significance, or produce an incorrect analytical conclusion.

Publication by CHRONOS ALERTS does not guarantee that every fact, interpretation, score or conclusion is correct. Review the cited evidence and independently verify information before making consequential decisions.

When credible new evidence materially changes the factual record, or shows that CHRONOS ALERTS was wrong or incomplete, CHRONOS ALERTS may correct, update, resolve or retract an alert. Material revision history is preserved whenever technically available, so readers can distinguish earlier publications from later changes.

Multiple sources increase confidence but do not create certainty. Several publications may repeat the same original report or official statement, so source count alone is never treated as independent confirmation.

CHRONOS ALERTS weighs source independence, primary evidence, authority, timeliness, corroboration, evidence quality and known uncertainty.

CHRONOS ALERTS uses automated systems, including artificial intelligence, to assist with monitoring, source review, classification, summarization, comparison and analytical organization. Automated systems can make mistakes.

An alert is not represented as manually reviewed by a human unless that review actually occurred. Automated analysis does not transform a third-party claim into a fact confirmed by CHRONOS ALERTS. Every alert is designed so readers can inspect its underlying sources and stated uncertainty.

Alert addresses stay stable when titles or analysis change. Material updates, resolutions and retractions are preserved as numbered revisions whenever technically available. Alert pages show that available history so readers can distinguish earlier publications from later changes.