The Corroboration Record: How Long Crypto News Takes to Be Confirmed

Last checked against source: 19 August 2026

Crypto trade press has a confirmation problem that everybody discusses and nobody measures. A claim appears, it is repeated within the hour by outlets that have added no reporting of their own, and by the time a second organisation has actually checked anything the story is already established. We built a newsroom that counts independent publishers per story, which means we can put numbers on it.

How long a story takes to be independently confirmed

Measured over the last 14 days, across the 1,259 stories that reached a second independent publisher at all. The clock starts when the first tracked publisher carries the claim and stops when a different one does.

The median matters more than the average here, and it is the number worth carrying away: half of the stories that eventually get confirmed take longer than 2.5 hours to get there, and roughly a third are still resting on a single outlet six hours after they first appear. Anyone reading crypto coverage inside that window is reading something nobody else has checked.

Our own weekly record

Stories we published, grouped by how many independent publishers ever carried the same claim. This includes our first weeks, which were bad.

The first row is not a good week. In the week beginning 2026-08-03, 68.8% of what we published rested on a single source. We were publishing far too much, too fast, and the corroboration count was a label rather than a gate. Cutting the volume and making the gate binding moved that to 1.7% by the week beginning 2026-08-17. We are showing the bad number because a quality record that only contains good weeks is an advertisement.

Method

Stated in enough detail to be checked or disputed.

• What counts as a publisher. A source in our tracked list with the corroboration flag set. Aggregators and syndication feeds that republish without adding reporting are tracked but excluded from the count, which is why the corroborating set (141) is smaller than the tracked set (166).
• What counts as the same story. Reports are grouped into clusters by the pipeline. Two reports in one cluster are treated as the same claim; the publisher count is the number of distinct corroborating sources in that cluster.
• When the clock starts. The earliest publication timestamp among the first publisher’s items, falling back to our fetch time where a feed omits one.
• What is excluded. Our own briefings, because we wrote them and counting ourselves would be circular. Unpublished and held stories are excluded too.

What these numbers do not say

• They are not a survey of crypto media. They describe the sources we track and the stories we clustered. A different source list would produce different numbers.
• Unconfirmed is not false. A story resting on one outlet may be entirely true and simply exclusive. The count measures independent verification, not accuracy.
• Timestamps are imperfect. Feeds report publication times inconsistently and some backdate. Where a feed gives no timestamp we use our fetch time, which can only make a gap look shorter than it was.
• The time-to-confirmation figures exclude stories never confirmed at all. Those have no second timestamp to measure, so including them would require inventing one. The weekly table above is where they appear.

Using this data

You are welcome to cite, quote or reproduce any figure on this page, including the ones that reflect badly on us, with attribution to AltcoinGordon and a link to this page. If you want the underlying weekly series in a machine-readable form, or a cut we have not published, ask at contact and we will send it. If you think the method is wrong, we would rather hear it than not.

Related: how the publisher count works on an individual story, and our editorial standards.