I wonder if you could explain the HN behaviour I describe at:
https://news.ycombinator.com/item?id=38608546
I'm the author of that submission and the blog post it points to. I submitted this item immediately after publishing the blog post I had written. Nobody could have, or did, submitted it earlier. Anyway, it didn't generate much interest on HN.
Later that same day somebody else submitted it and that time it hit new and got some good interest. Cool!
Q1: how could the other person post the same link I had already posted? I thought duplicates were prevented?
A further day later and my original post has a more recent submission date. And as it gains interest it is marked as a dupe.
Q2: Why is this original submission of mine now showing a more recent date? Is that because somebody submitted a duplicate that was caught?
Thanks!
> Why is this original submission of mine now showing a more recent date?
I think it's got a second chance, part of the 'pool' https://news.ycombinator.com/lists I see it on the 'pool' list but I'm not one of the inner circle having special buttons (I'm just guessing it works with buttons or links somehow).
He has also suggested to me to hang back and repost 10+ days later for stories he thinks had merit. (I hasten to add this has happened precisely once over the years. I am not show pony here, I was lucky finding a thing which he thought genuinely didn't get the eyeball time it deserved)
I am unsure there is an algorithm here. I tend to believe its a heuristic, there is some formulaic weighting, but also some prodding with a stick.
On HN a duplicate is either a submission of the same link, or a substantively similar article which doesn't advance significant new information on a story.
Where multiple submissions occur in a short period of time (say for a major breaking story), HN typically aggregates discussion under the first submission, usually with a note. Votes may or may not be aggregated (and often aren't), and this can sometimes have an influence on the "flameware" detector which is based on the ratio of comments to story votes (> 40 votes, comments > votes), which I've suggested may unfairly penalise some high-interest / high-impact stories.
In general, if an item's been posted within the past running year and there's been more than about 20 comments on the thread, any substantively similar submissions will be considered "duplicate".
(Dang / HN mods can also see residence time on the front page, which may enter into their assessments. Ordinary participants cannot see this.)
Submissions without substantial discussion aren't dupes.
Submissions occurring over a running year after the most recent substantive discussion are not a dupe.
And for large and complex stories (SBF, OpenAI, the Reddit boycott, long-running legal battles, COVID-19, BLM, etc.), new stories which introduce significant new facts are not considered dupes.
(Additionally HN may up- or down-rank specific items based on source, various "hot topic" items, etc. There are many dang comments discussing this.)
The 2nd chance, invited post, and a few other pools may occasionally create disruptions in the first-past-the-post spacetime fabric, as dang notes occurred here, in this comment:
<https://news.ycombinator.com/item?id=38633974>
My view is that exposure and discussion of a story, whether I or someone else has submitted it, is the chief goal. HN karma's a mostly irrelevant score, and there are numerous high-quality contributors and industry idols / domain experts with very low karma, as well as space alien cats with too much time on their paws with far more than they deserve.
In other words, sometimes second-chance picks end up duplicating a later discussion. We try to avoid this but it happens!