Methodology

Severity measures disruption to the movement of goods, not human cost. The highest level whose criterion is met by evidence in the record applies. A change in level after publication is logged as a correction.

For each affected port or chokepoint we compare the daily series during the event window with what the same weekdays looked like in a 56 day baseline that ends 8 days before the event starts. The 7 day gap keeps pre-event rushes out of the baseline. Days inside other recorded events and national holidays, including Lunar New Year, are excluded.

We report the expected daily average, the observed daily average, the absolute and percent change, and a standardised score: the average daily deviation measured in robust baseline standard deviations. Every figure appears with both windows, the number of valid days, the method version and the date it was computed. When upstream data is revised, effects are recomputed and earlier values stay visible.

This is a measured change in public data during the event window, compared with a baseline before it. It shows timing, not cause. We describe an event as causing a change only when an official source says so.

The full method, with the floors for small counts, the holiday calendar, the year-ago check and how to reproduce every number from published files, is on the effect method page.

Software flags candidates in two ways: a statistical check when a series moves 3 or more baseline deviations for 3 nights running, and matching of official notices and trade press headlines. Candidates go to a review queue. A person checks sources, writes the record and approves it. Nothing reaches the archive without that approval, and the site build refuses any record without it.

• Summaries and timelines are in our own words. Quotations are official statements only, under 15 words, attributed.
• Every factual claim cites a source; every official and press source is archived.
• No speculation about cause. Disputed or investigated causes say so, and say who is investigating.
• Casualties appear once, from official sources only.
• No private individuals are named.
• Corrections are dated and shown on the event page and the corrections log.
• No commercial freight rate index is ingested, stored, charted or quoted as a value. See sources.

Daily port calls, transit counts and trade volume estimates come from IMF PortWatch, which derives them from AIS vessel position data. Vessels with transponders switched off are missed, and volume figures are modelled from vessel draft and capacity rather than customs records. PortWatch rebuilds its tables weekly, so past values can change; we keep every revision.

These read the same approved records and public series from a different angle. None of them is a new source of facts, and each says so on its own page.

• Ongoing disruptions: a live view of every event marked ongoing, day by day.
• Recovery time: for ended events with a measured, material fall, how long the series took to get back to about its pre-event level.
• Trade measures tracker: sanctions, customs and capacity-withdrawal events, split into in force and ended.
• Confidence audit: every record's own stated confidence level, in one list.

What counts as a freight disruption event?

An event is a dated disruption to the movement of goods at a port, terminal, chokepoint or trade lane, such as a strike, closure, grounding, canal restriction or cyber incident. It gets a record only when official or press sources document it and an editor approves the record.

Does a measured effect prove that the event caused the change?

No. A measured effect shows how a series changed during the event window compared with a baseline before it, which is timing, not cause. We say an event caused a change only when an official source attributes it.

Where do the port call and transit numbers come from?

Daily port calls, chokepoint transits and trade volume estimates come from IMF PortWatch, which derives them from AIS vessel position data. Vessels with transponders switched off are missed and volumes are modelled, so small changes need care.

Why does this site not publish freight rates?

Freight rate benchmarks are licensed commercial products, so we name and link them but never store, chart or quote their values. Where a price direction is needed we use the public domain US producer price index for deep sea freight and label it as such.

How is severity assigned?

Severity runs from level 1, minor, to level 5, critical, and measures disruption to the movement of goods rather than human cost. The highest level whose criterion the evidence in the record meets applies, and any later change is logged as a correction.

How do I report an error?

Email corrections@worldfreightmonitor.com with the page address and what is wrong. Every accepted correction is dated and shown on the record and on the site-wide corrections log.

What is the confidence level on a record?

Confidence is the approving editor's own judgement of how well sourced a record is: high, medium or low, with a written reason. It is set once, at approval, and not recomputed. Every record's level and reason are listed together on the confidence audit page.

How is recovery time worked out?

For an ended event with a measured, material fall (a standardised score at or below -2, the same bar severity level 2 uses), recovery time counts the days from the end of the measured event window to the first 7 day trailing average that reaches 90% of the pre-event baseline. It is a plain reading of the same public series the effect method uses, not a method figure: no baseline exclusions, no weekday medians, no statistical test, and an event that has not reached that level by the latest data reads "not yet", never "never".

What is on the trade measures tracker?

A view over the archive's own event records, not a separate database: every published event of type sanctions action, regulatory or customs action, or capacity withdrawal, split into those still in force and those ended or unconfirmed. A militarily enforced closure, such as a canal blockade, is filed as a security incident and appears among ongoing disruptions instead.

What is the weekly digest on the home page?

A short, automatically built summary of the last 7 days of data: new events, corrections, and the biggest 28 versus prior 28 day moves among ongoing disruptions’ own measured series. It only reflects records already published elsewhere on the site; it adds no new judgement of its own.