It's worse, even ... there was also a threadreader (unrolled) thread (#90304) as well.

The first problem, dup detection being case sensitive, was reported as Issue #142.

The next problem is how if you try to Search for a link (e.g., to see if it has already been posted), the search results do not include Link Posts where that link was used. I just reported this, as Issue #212.

There's also the problem where one person can post the mobile.twitter.com/status/nnnnn link and the dup detection will not know that is the same as twitter.com/status/nnnnn (i.e., without the mobile. subdomain). That's reported as Issue #208.

Similarly, www. subdomain has the same issue, and that was reported with Issue #169.

There's also the problem of URL shorteners, where a Link post could be using the T.co (shortened) link that Twitter uses in the body of a Tweet. So that's addressed in Issue #179

And there's a delay in detecting a dup even, so a person composing a Link post who pastes the link and is quick at hitting Post, ... the dup detection might not have been finished, and the user won't even see the "dup" warning. This is reported as Issue #153

And there are a few other Issues reported for lessening the chances of a dup occurring (e.g., when the link uses Google AMP, etc.).

So addressing these will help lessen the number of occurrences of accidental dups.

Not all dups are accidental. I suspect some people see the Dup, and know it is a dup, and post it anyways. Maybe because they want to earn sats? Maybe because the earlier post was days ago (or longer) and want to promote it. Lots of reasons could explain it. And then you have the ones who create a Discussion post and then in the discussion include the Link, (and as a result, the dup detection won't even be attempted).