Hacker News Reader

Comment by philippnagel | Hacker News Reader

How will "most interest" be defined?

Seconded. Additionally, I'd love to see a distribution of comments/upvotes by date of submission (such data I think could be gleaned from HN search). My eyeballs tell me posts submitted within the first 2 or 3 days of Apply HN got much more traffic/discussion. Certainly it is possible that those first posts warranted such discussion and subsequent posts did not. Or maybe I'm completely wrong in my observation, which is why I'd like to see such data.

Also, after ApplyHN became randomized (around midway thru the experiment) it seemed like discussion dramatically fell again, but it was definitely tough to tell since refreshing showed new applications every time.

Fun experiment though and I'm excited to see what types of projects the community picks.

Yeah, what was the criteria for making it to the main Apply listing? Was it number of votes? Comments? Our application never showed up there: https://news.ycombinator.com/item?id=11467813

Yep, mine never made it to the front of ApplyHN either. It seemed like unless it got at least one or two immediate upvotes, it disappeared into the abyss, thus making it even less likely to get traffic.

I second this. After the first 2-3 days there was a sharp drop in engagement in general, though there were some applications that did break through.

When ranked by upvotes, of the top 5 applications, 2 were posted after the first 3 days.(https://medium.com/@ApplyHN/top-5-applyhn-submissions-ranked...)

And 6 applications posted in the first 3 days reach top 10 and only 4 from the remaining 18 days.

Upvotes * Comments ? Of course that wouldn't be great for one where everyone thought, "yup, that's perfect", upvoted and didn't say much. Upvotes + Comments. Maybe top 10 upvotes, top 10 comments, top 10 multiplied and remove duplicates?

Edit: or maybe the cases where "everyone thought" "yup, that's perfect", upvote and no comments is an indication of fraud?

Many comments may mean that the submission is good, or it may mean that the submission is bad and the comments try to show the problems. Usually lot of comments mean more "polemic" than "good".

Good point. There should definitely be a manual culling of the set of 30 (besides duplicate removal). To make sure there aren't just a bunch of people saying "that idea sucks!" and that there isn't a vote rigged entry (a poorly presented, low commented proposal that magically has a bunch of upvotes).

He made it clear in the last post it would not be literally most votes in the case of someone beating their fraud detection.

Beyond passing fraud detection, someone with a large, strong social network could get a lot of votes. Is that a good signal for a startup? Maybe it is, maybe it isn't.

Reply on news.ycombinator.com