Why we revert to original titles
There is an ongoing trickle of complaints about this, as if we were engaged in some sort of sinister conspiracy.
Titles on HN are not self-expression the way comments are. Titles are common property. The person who happens to submit something first shouldn't thereby get the right to choose the title for everyone else. This would be clearer if we didn't let submitters enter a title-- if our software simply let people submit urls, and retrieved the title from the page. We don't do this because it's too inflexible. Some articles have titles that are too long. In others the subtitle makes a better title. But the fact that a title field is editable doesn't make it comment.
It's true that when submitters change titles, their new titles often contain more information than the article's original title. But a significant percentage of the extra information added in this way is false. The only way we can tell if a newly created title is accurate is to read the article, and we're not about to read every article submitted to HN. The only option is to revert to the original title, which is at least what the author intended.
(We do sometimes change titles from the original when the original title is egregious linkbait, or false. We have also, since the beginning when our users were largely YC alumni, put e.g. (YC S13) after the names of YC companies in titles. But these are not the types of changes users mean when they complain about moderators changing titles.)
If we had infinite attention to spend on moderation, we could read every article and decide whether each user-created title was better than the original title. But we don't. Moderating HN is no one's full time job. And frankly it's not that big a deal anyway. If we're going to expend cycles trying to fix something about HN, the increasing prevalence of mean and stupid comments has a much higher priority than the fact that authors' original titles are not maximally informative.