I'm quite familiar with Betteridge. Strict adherence gets to your question; the larger issue, as with anything, is that so many question heds have been posed where the answer is "no" that it became a "law."

In journalism, there are problems that showed up far earlier than clickbait question heds, such as garden-path or irrelevant ledes. I've written and run question heds that were correct display copy atop stories in which the reporter tried to find an answer but couldn't given conflicting information from sources. At that point, the correct approach is a question.

Question heds atop stories that definitively disprove the question are lazy at best and disingenuous at worst. But to categorically remove a form of hed writing as valid based on statistics or anecdotal data isn't an improvement.