I remember such situation in an online multiplayer game.
The game had servers associated with different sites providing login. Of course, the largest server was the one associated with Facebook. But we're not talking about this server; instead, we're talking about a much smaller server associated with a smaller site. This server was clearly a "second-class citizen" in the view of the developers (likely because of its humbler popularity) and therefore, while it theoretically did have moderators, they were disinclined to burden themselves with actual moderation and it was nigh-impossible to make them step in and take any action.
Usually, there was no need to. There was a chatroom in this game; and in spite of the moderators' indolence it did not detoriate into total chaos. There was one particular user who became a de-facto informal moderator, and I must admit, she was doing it well. Whenever someone was becoming excessively unreasonable or disruptive, she would typically say she'd muted them; and this served as an informal global mute, since at this point most other users would either mute this person themselves or, at least, ignore and stop responding. She was also very good in tactifully telling other users to improve their behavior whenever need was arising. But, while she was helpful, she wasn't necessary; even when she was offline, other established users were able to maintain the chatroom in a healthy state. Effectively, the community was self-sufficient.
There was one exception. Occasionally, a disruptive user was dropping in and was starting to spam the chatroom with their, sometimes vulgar, garbage. The aforementioned measures were not working since he was not trying to be a member of this community; rather, he was simply dropping in, spamming for some time, when most people had muted him he was dropping out, waiting till the mutes had expired and then appearing again. It was annoying and this was the only situation I'm aware of when his community was trying to reach its official moderators, asking them to ban him; but, as I said, they were unreachable. They wouldn't bother to lay down one single ban. Still, while this was mildly annoying, this wasn't enough to seriously damage the community.
I'm wondering, if such a self-sufficient community is the optimal model whenever it actually works? And under what conditions can it actually work? Usually, my intuition was that a non-moderated chatroom or forum or whatever would likely quickly be overtaken by low-quality content and trolls and would soon detoriate to what can be seen in chan-related parts of the Internet. Yet, it was not the case. Under what conditions can an online community (with free registration) become self-sufficient in that way?
If it seems the community can become self-sufficient, is it optimal if the moderators simply let it go and refrain from stepping in, unless a spammer appears? (Because I'm pretty sure that not bothering to ban this spammer was far from optimal, wasn't it?)