Reviewed 30 September 2026 · quarterly cycle
How do I check a whole saved list for dead links?
Sample rather than sweep. Ten addresses taken at random tell you the health of the list in ninety seconds, and checking all of them tells you the same thing after an hour.
The evidence
The reason sampling works here is that address decay is not random. Entries decay in the order they were added, so a list decays from the oldest entries forward. A random sample of ten will therefore land on a mix of old and new entries, and the failure rate in the sample tracks the failure rate in the list closely enough for a decision.
Where sampling misleads is on a freshly rebuilt list, where the oldest entries have been replaced and the decay is concentrated in a band rather than spread. On a rebuilt list the more informative test is to check the ten oldest entries rather than ten at random, because that is where the rebuild left the accumulated age.
The practical procedure is three steps. Resolve each sampled address through to its destination rather than stopping at the first response. Compare the item title against the entry title, because a resolving address with a different title is a renewal rather than a live entry. Record the failure count, and treat anything above a quarter as a signal that the list needs a rebuild rather than a spot fix.
When this does not apply
Where a list uses short addresses throughout, resolution is the only available test and the item title comparison becomes the whole check. Lists in that shape take longer to test and produce a less clear result, which is a reason to prefer lists that publish full addresses.