Here’s why Paul Graham (probably) owes me an apology.

What I am about to say here is not based on a hunch. I don’t say this kind of stuff lightly, and I will give evidence.

Additionally, I make no pretense of knowing whether Graham himself participated in the adverse action. I know that people other than him have been given editorial power, and that some have, in the past, displayed serious incompetence, especially with regard to banning of posters. Reckless silent banning (“hellbanning”) is a notorious problem on Hacker News, and the general consensus is that the problem is with Graham’s hired editors, and not the man himself.

However, I believe that I am owed an apology. (If not, I seriously misunderstand the Hacker News comment ranking algorithm, which I believe to be based on age and votes of the comment.) In September, I was tipped-off (anonymously) that my comments were being downgraded in determining their position on the board. I’d suspected this, since I went at one point from several top comments to seeing mine often at the bottom, despite several upvotes.

At first, I didn’t believe that I’d been personally targeted because, while I have been critical of the VC-funded ecosystem (“VC-istan”) I’ve always supported Paul Graham and what he’s been doing. Additionally, I’m a high-quality contributor to the board: top-100 karma, high average karma, well-written comments. However, there did seem to be a change in the ranking algorithm; something flaky was happening. The tipper gave me a way to test whether something wrong was happening: check latency times from Hacker News while logged in, and though an incognito window (“slowbanning”). Sure enough, the 3- to 5-second latencies I’d experienced under my personal account were not experienced in incognito mode. Typical latency for me, when not logged in, is 250-400 ms.

A few years ago, Hacker News made karma ratings invisible. At the time, as I recall, the justification was to prevent runaway behavior whereby popular comments got even more (less deserved) upvotes than ones with mediocre scores. I now tend to think that the change may have been made for a less noble reason: to hide that certain posters were getting (in terms of comment placement) personal penalties. If it were true (and it may not be, and if I am wrong, then I will be the one owing an apology) then it would be extremely damaging to morale, and hiding the fact would be mandatory.

If I were being personally penalized (presumably, in direct retaliation for my criticism of the VC-funded ecosystem) it would be visible by looking at older comments. Why? With HNSearch, one can find the karma rating of any comment older than about a week. (You can, if you wish, check the assertions I am making here.) Additionally, multiple sources have given me that the age penalty on comments is on the order of (2 + t)^-p, where t is the comment’s age (in hours) and p is an exponent variously given between 1.5 and 2; this means that for old threads (t >> 1000 hours) age would, assuming the comments were within days of each other, which they almost always are, be pretty well factored out.

Looking into some high-karma comments, such as 99-point comment 5 months ago, I found strong evidence of a personal penalty. That comment was placed near the bottom, while the top comments on the thread had substantially lower scores. This is not of direct concern to me (few people read 5-month-old threads) but it confirms what I had suspected.

There are 3 possibilities.

  1. (Least likely.) In retaliation for my criticism of the VC-funded ecosystem, Paul Graham has assigned a personal penalty to my comments, causing them to fall to the bottom, even when they are highly rated. If this is the case, I am owed a personal apology by Paul Graham.
  2. (Most likely.) My comments have been personally targeted, but Paul Graham is not the culprit. In this case, Paul Graham should apologize not only to me, but to Hacker News at large, for giving editorial privileges to the incompetent who pissed off a top contributor. Additionally, this person should be dismissed from the editorial role.
  3. (Other.) There’s something I am missing: perhaps an additional complexity in the ranking algorithm not related to comment success (karma) or age– I’ve tested for both. I can’t see what, but I’d be open to an explanation outside of what I’ve covered here. If I am wrong and my comments are not being penalized, then I must offer my apology for the suspicion. It is not one that I developed (much less voiced) lightly, but I cannot for certain guarantee that I am not wrong.

I look forward to learning which of these is the case. If I am not alone in suspecting chicanery, it should be discussed.

13 thoughts on “Here’s why Paul Graham (probably) owes me an apology.

  1. It’s possible and likely that you’ve received a lot of “flags” over your lifetime. The algorithm may take your history into account when ordering your comments, even those without controversy. This would mean it’s not personal and that nobody even considered you individually when making the decision that had this result.

    • I agree with this comment. This alone could be enough. Your comments are very controversial, and tend to take a negative tone more often than not. That alone attracts comment flagging. I’ve been learning to not take things personally as time goes on and I find out things that I first thought was personal was just a widespread policy, my own mistake, incompetence or even just imagined.

    • Agree. Remember that being flagged on Hacker News doesn’t require pg to do anything personally — that’s an “other readers” thing. Your VC-istan posts may not bug pg — most probably hypothesis? he’s never read them and has no opinion. But I’m *sure* they bug somebody who reads Hacker News, so you could still be getting VC-istan downvotes even if they’re primarily from random readers.

      Indeed, that’s far more likely.

  2. For what it’s worth, I think your writing is more interesting than Graham’s, and whatever’s going on here, you should probably just ignore it.

  3. Hacker News has jumped the shark a long time ago, and if your assumption is true then it’s just further evidence of its decline. Your comments are, in my opinion, among the most interesting and insightful on that site. However, I can also see why all those startup cheerleaders and web developers would feel offended by your oftentimes brutal honesty. At least you didn’t get “hellbanned” yet, which has happened to quite a few other excellent posters on Hacker News who chose unpopular truths over groupthink.

  4. I like your writing and your ideas, but they are so angry and spiteful. I was so happy for you when you once wrote that you were quitting Hacker News. It seemed like you had learned that arguing in that forum was a huge drain on you. I imagined that your internet persona would go dark, and you would find something for yourself in life. It looks like you went back. I hope that you take down this post, too.

    • Angry, yes; spiteful, no (at least not here on the blog; I don’t read HN). But Michael, I agree that, given your compelling reasons for leaving that wretched hive, you shouldn’t have let yourself be drawn back into it: it only leads to headache and heartache.

  5. Pingback: The difference between unfairness and injustice, and why it matters | Michael O. Church

  6. Unless HN was some sort of a critical component of your advancement, career-wise or intellectually – why do you care?

    • Honestly, it doesn’t bother me that much. It’s just annoying because I put a lot of thought into what I have to say, and I shouldn’t suffer a personal penalty just because some rich guy is afraid of me. If my comments fell to the bottom because the HN community found them less useful (i.e. they were downvoted) then I would not have a problem with that.

      I’m not so much upset as desirous of drawing attention to this. Why? Because it means that some very influential people are scared of me. It means that what I have to say *matters*. People should know that.

  7. Personally, I think you’ve hit the flamewar detector. I suspect (and have no evidence) that given that there is a flamewar detector, and that its purpose is to prevent flamewars, then it may (as well as dropping articles down massively) penalise the comments of those who have been involved in multiple flamewars. The idea would be that if certain users tend to be involved in flamewars, then by penalising their comments, flamewars become less likely to happen. Just my two cents (but I do have a hobby of attempting to induce algorithms from observed consequences).

  8. Pingback: The Rise of Hacker News Is Bad News for the Tech World ← Dennis Forbes...Professional

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s