[dang]

What you observed is so within the normal range of HN's functioning that there's a FAQ entry dedicated to it:

  Q: Why is A ranked below B even though A has more points and is newer?

  A: You can't derive rank from votes and time alone. See "How are stories ranked?" above.

https://news.ycombinator.com/newsfaq.html

In this case, the post dropped in rank because of user flags.

[fakedang]

Understood. Is the flagging mechanism based on relative counts or absolute numbers? I think the latter mechanism would be extremely flawed.

[dang]

I don't understand your question - can you put it differently?

[fakedang]

So let's say we have a post with 110 upvotes and 40 comments, as well as 30 flags (i.e. 150 post interactions and 30 flagdowns, or a 1:5 ratio of flagdowns to interactions), and another with 27 upvotes, 13 comments and 20 flags (i.e. 40 post interactions with 20 flags, or a 1:2 ratio of flagdowns to interactions).

The second one is clearly more relatively flagged down than interacted with, even though it was flagged less. This could mean that the topic was clearly against the guidelines for glaring issues. On the other hand, the first could very likely be a controversial topic of discussion with clearly redlined sides (like the recent Jacobin mag article on men's health), or (unlikely but possible) that a swarm of company Z employees could have zealously flagged down the post because of its critical nature against company Z.

In an absolute flagging system, the first article would be flagged, by virtue of having the higher number of flags. In a relative flagging system, the second article would be flagged.

I assume that you hold final authority on which flagged posts actually get flagged and which stay. I hope it's not another decision that's been relegated to an algorithm in the background.

[dang]

Ok, in those terms, HN's flagging system is relative, not absolute.