Hacker News Reader

Comment by tptacek | Hacker News Reader

They're entirely forthcoming about this: except for job and launch posts, YC posts aren't ranked differently, and YC company upvotes are treated no differently than anyone else's. I think it'd be hard to count how often Dan has said this.

You can choose, of course, not to believe him about this (I have a hard time thinking of a person in this industry I trust more than Dan), but you can't ask him to be more forthcoming about it!

View on HN

Sure, I also, still, feel some of the kinetics need to be explained to avoid confusion.

Like the homepage doesnt make much sense to me after all these years of reading HN. As I write this there are 160 upvotes in one hour on this post and it is still #9, with many weaker posts ahead of it.

The general answer to this is in the FAQ: https://news.ycombinator.com/newsfaq.html.

The specific answer (about this post) is that moderators haven't touched it in any way. Twitter.com links have a mild downweight, along with many other domains that produce a lot of sensational/offtopic content (relative to HN). That includes nearly all major media, forums like reddit, etc. Other than that, all you're seeing is the regular HN algorithm at work. It's more complex than just points + time—we've tweaked it over the years to (try to) do a better job of selecting for intellectual curiosity, dampening flamewars, and so on.

It's been known that HN has applied keyword-based, domain-based, and manual penalties in the past: http://www.righto.com/2013/11/how-hacker-news-ranking-really...

While I doubt there is a "non-YC penalty", and I doubt there is collusion on dang's part, it isn't correct to think of HN ranking as a unmoderated scoring function.

I didn't say HN operates on an "unmoderated scoring function"; it most certainly doesn't, as Dan has (again) repeatedly said. I said that YC posts aren't ranked differently, and YC votes don't count differently.

You are arguing the specifics of my comment and not the spirit of it. It’s not clear all the ways posts are modified and it isn’t clear who all knows all the ways posts are identified.

It's clear enough, if you pay attention. Follow Dan's comments over time; they're an informal moderation log of the site. You'll see what stuff does and doesn't happen to posts.

My comment was more about what are all the different ways post can be ranked higher/lower. I trust Dan that the two things I mentioned they don’t do, but is there anything else like keywords/hosts that if yc members knew would give them an edge.

There is no reason to believe anything like that is happening. We do well enough on HN that Dan called us out downthread, and certainly there are no secret tricks we're using --- short of "not having our friends vote us up when stuff of ours gets submitted". I think this is pretty much just conspiracy theories.

The discussion here is pretty unproductive. You're free to simply disbelieve Dan; that's a coherent perspective to have, even if I think it's deeply wrongheaded. But I assume we both know enough about the dynamics of message board "debate" to know that there's an unbounded set of un-falsifiable arguments you can throw up to support the idea that Stripe is unfairly gaming HN. Maybe they've found a wrinkle in the way the second-chance pool is handled!

You are quite defensive and I’ve never said I didn’t trust Dan but you keep saying it.

Because you knew enough about HN that this conversation is “unproductive” doesn’t mean it is unproductive to others. My comment got a lot of upvotes. Maybe that means the knowledge you have isn’t as wide spread as you think it is.

Also I recommend you compare Dan’s responses in this thread to your own. A very different tone—one that is “more productive” than how you’ve chosen to respond.

My entire point was more information could a. Prevent these posts in the future and b. Help to others understand the dynamics when a post like this is made. What is wrong with this? Jeez... was my initial comment that offensive to you?

Reply on news.ycombinator.com