Your analogy adds nothing but noise to the conversation.
The lawyer for the accused never recants either the confession, or that his client said what he is reported to have said. In typical defense lawyer posturing, instead he tries to draw doubt about the metadata around it. And for good reason given that the accused has repeatedly admitted, to the police, prosecutors, and even the media, having written exactly what was reported, but now we're to pretend that maybe it was a different person, or that Facebook messages were doctored?
Note that the prosecutor almost certainly has the messages from Facebook, and the fact that the police went on the early variant is utterly irrelevant.
What interest does the law have in punishing people who aren't actually issuing threats?
Since you like anecdotes, imagine that someone walks around pretending to punch random bystanders, but he pulls up at the last second. Hilarious, right? Now they have no intention of actually hitting, and can even point at a history of being the sort of person that pretends to hit people.
Is it assault?
Yes, of course it is. The recipients of the threat reasonably believed in its validity, so what the perpetrator thought is irrelevant. This is the same reason why bomb threats, threats against persons, etc, are prosecuted, despite this contrived notion throughout this thread that they need to prove that one has the means and intention of carrying it out.
Freedom of speech is one freedom. But as with most freedoms that need to be balanced between people, your freedoms end where my nose begins. Your freedoms to express violent threats ends where there are people who might reasonably assume them to be legitimate, for instance. This is the case in the US and almost all Western countries, so it's odd seeing so many seemingly thinking this is protected speech. It isn't.