The design element on this site people compliment most is the rubber stamp. The design element that actually matters is the four-digit number inside it.
Every fabricated headline we publish is assigned the next number in a single sequence. The number is unique, it is never reused, and it is never skipped. The archive lists all of them. That means the integrity check on this entire publication is available to a stranger with no access to our systems: read the archive, sort by number, look for a gap. If we ever quietly withdrew something inconvenient, the hole would be sitting there in the sequence with nothing in it.
This is not a novel idea. It is roughly how error logs, court dockets, and check registers have worked for a very long time, and for the same reason: a sequence you cannot edit without leaving a mark is worth more than an assurance you cannot verify at all.
Anyone can promise to be accurate. A number that would visibly be missing is a promise you can audit without our help.
What makes it a promise rather than a decoration is the part we do not enjoy. Some headlines go up before their correction is written. We could hold those back and publish nothing until each one had its article finished, and the site would look immaculate. Instead those cards are marked correction pending, they are deliberately not clickable, and the footer counts them in a separate tally that sits next to the other two. A reader can see the gap between what we have promised and what we have delivered at any moment, without asking us.
We are aware of the obvious criticism: that this is accountability theater, and that a publication which invents its own scoring system will always score well. Fair. So here is the test we would apply to anyone claiming what we are claiming, ourselves included.
First: is the record complete, or only the flattering part of it? A corrections page that lists corrections but not the claims that needed correcting is a highlight reel. Second: can the count be checked by someone outside the building? If verifying the claim requires trusting the claimant, no claim has been made. Third: does the mechanism cost the publisher anything when it fires? A process that only ever produces good news about the people running it is not a process.
We think we pass those three. We will say plainly that passing them is not the same as being right about anything. A contiguous correction ledger tells you we have not hidden a retraction. It does not tell you our articles are good, our sourcing is deep, or our judgment is sound. Those are separate questions and the number does not answer them.
But it answers one question completely, and that is more than most publications can say about any single element on their page. The number is the smallest unit of trust we know how to ship. Everything else here — the ticker, the stamp, the labeled images, the fact that the fake quote is always attributed to someone we openly invented — is an elaboration on it.
Take the number away and this is a joke site with good typography. Leave it in and it is a joke site you can audit.