Revert the same article four times within twenty-four hours on English Wikipedia and a block follows almost automatically, whatever the fourth edit actually restored. The mechanism has a name, WP:3RR, and a specific shape: not four reverts in a calendar day, but four in any rolling twenty-four-hour window, timed from the first revert to the last. Cross that line and the content dispute stops mattering. The question an administrator asks isn't who was right. It's whether the count reached four.
"Revert" is defined more broadly than the undo button. Restoring a paragraph someone removed counts. Removing a paragraph someone restored counts. Rewriting a sentence back toward an earlier version, in different words, counts. The handful of exemptions — obvious vandalism, a living-person violation, a clear copyright problem — don't reach an ordinary sourcing disagreement, however confident either side is that they're the one being reasonable. And the bright line at four isn't a permission slip for three: the broader edit-warring policy, WP:EW, covers a sustained pattern of reverting at any count, if that's what the page's history shows.
Picture a company account watching a paragraph about a lawsuit settlement appear in its own article, sourced to a regional business paper. The account removes it. Another editor restores it. The account removes it again, then a third time, each edit summary a little more emphatic about the source being unreliable or the detail being disproportionate. By the third revert, the page is usually more visible to other editors than the account intended — on someone's watchlist, or flagged by a bot that tracks revert counts. What happens next is rarely a ruling on the source. An administrator blocks the account, or protects the page at whatever version happened to be live when they looked, sourcing question still open, now frozen in place until someone reopens it properly.
A block doesn't settle who was right about the source. It just decides who gets to keep arguing.
That freeze has a name in Wikipedia's own culture — editors call it getting stuck on "the wrong version," because full protection locks whatever text was on the page at the moment an administrator intervened, not whatever text either side thinks should be there. It's a joke with a real point: protection is a pause button, not a verdict, and the account that triggered it has generally lost more than it gained. A conduct problem now sits on top of the original content problem, and conduct problems are harder to walk back than sourcing disagreements.
For an account with a disclosed connection to the subject, the cost compounds further. Wikipedia's conflict-of-interest guideline already discourages editing the article directly; a page representative who reverts three times to protect the company's preferred version has supplied exactly the pattern a conflict-of-interest noticeboard thread goes looking for — not an isolated edit, but a fight to keep control of the text. What began as a question about one source can turn into a question about the account itself.
The alternative costs one revert, not three. Remove or restore an edit once, then move the argument to the talk page, where it belongs: cite the specific line in the sourcing guideline the removed content fails, or the specific line it satisfies, and ask. If nobody responds, a formal edit-request template or a post to the reliable-sources noticeboard keeps the question open in a venue with a record other editors can act on. None of that guarantees the paragraph stays out, or comes back in. It does keep the argument about the source, which is the only argument anyone was ever going to win.
