How fake news spread on gaming forums

I wrote multiple posts about the game I currently play, World of Warships. Several of them (1, 2, 3) needed stupid amount of data processed to measure something that bothers the players a lot. The ranked play of the game – besides the usual “win and progress, lose and fall back” – contains a mechanic that rewards the best player of the losing team by not falling back on ranking. This is made to motivate players to keep fighting if their team is behind. According to a widespread belief in the community, there are a large group of exploiters who play in a way that optimize their chance to be first among losers at the expense of team win chance. So they progress when the team can carry them and don’t fall back when they are too much to carry.

So after processing available data, I concluded that there is no significant group with the signature of these exploiters (low win rate, high experience points which determine who is considered best). I linked my posts on the subreddit (1, 2, 3) of the game and got downvoted and trolled, because I questioned their beautiful conspiracy theory that “explained” why they can’t get good rank. Finally I was accused by mods that I just want to drive traffic to my blog for money, despite the blog doesn’t even have monetization. Then – when I attempted to reasonably discuss another game feature, I was banned from their subreddit for “being toxic”.

Instead of crying somewhere, I decided to teach them a lesson by creating a troll post (archive). I created a throwaway account and posted completely fake statistics that tell them what they want to hear: that exploiting is widespread and this is why they can’t get good rank. The plan was simple: let them upvote, celebrate and join the fake account in abusing me and then reveal the troll and make them look stupid. To make it more embarrassing, I placed multiple clues in the post:

  • The first letters of each paragraph in my text comment spelled “BULLSHIT”
  • I linked the original dataset which was actually the lyrics of the song “Never gonna give you up” by Rick Astley.
  • I put “Fake News” watermark to the right top corner of the chart
  • The scientific discussion made no sense even if the data was true: I was talking about “standard deviation of correlation” (which is nonsense, but also one upvoted comment that some random fool used to criticize my real post) and also talked about damage which wasn’t present on the chart
  • The post contained lots of abuse for the real me which is unscientific
    1. .
      Then something completely unexpected happened which turned some laughter at the expense of some morons and slackers into something much more valuable: some of my clues were found. Pretty soon someone checked the original data and later the “BULLSHIT” was found. But the obviously worthless and misleading post was not removed or at least downvoted and flamed, it rose to the top and it remained on the first page for a day:




      Its reddit statistics:
      107 (73% upvoted), 8.4k views. Also, my only comment (the fake discussion text) got 88 karma. Let’s compare it to my real posts about the same topic:

  • 0 karma (43% upvoted), 659 views
  • 0 karma (30% upvoted), 1.2kviews
  • 0 karma (16% upvoted), 977views
    1. Also, neither of my comments got positive karma. Not only my real posts got me negative karma, but their combined audience is about 1/3 of those who saw the fake post, despite it was caught very soon!

      This is how fake news spreads on social media and gaming forums: people support and spread lies they like, even when they are fully aware that they are fake while actively suppress – with the help of mods – the true information that contradicts their viewpoint.

Author: Gevlon

My blog:

18 thoughts on “How fake news spread on gaming forums”

  1. Reddit tends towards a strong hive-mind mentality in certain topics (just look at anything related to net neutrality), even more when the behavior is influenced by biased mods. But I take my hat off to you: That there is some master level trolling.


  2. Confirmation bias is nothing new.
    You got upvotes after revealing the scam cause people like a good joke.
    As for the ban: the mod in question (Vaexa) banned me for a week when I commented on somebody blaming his teams for losses. Somehow calling losers out is bannable.

    As for banning you, it is at least arguable. You made several posts on the same topic within 3 days, all linking to your personal blog (with possible financial returns). You are arguing the same point in all 3, you could just make it one post. Nevertheless, Vaexa seems to be banning people for any non-conforming statements. There is exactly zero value in the sub as the result – there are no good guides, game tips nor balance discussions. I go there for some funny game snippets only, everything else is just the same 5 memes and statements repeated ad nausea.

    In the end, your experience in this game is extremely short and one-sided. Consider the chance that your interpretation of statistics is wrong even if the numbers are churned out correctly.


  3. I was going to joke “because you post it” but damn, you actually do.
    Amazing. You’re beyond jokes.


  4. @stawek: I don’t need any experience of the game to process statistics. The numbers don’t lie, there is no significant “low winrate, high XP” group. What these guys really hate is that (like in practically all FPS games) camping and sniping is good play. The “Coward Hindenburg” who shoots HE from the back isn’t saving his star, he is playing to win and has better winrate and rank than those who flame him.


  5. You have shown that bad players are bad, whether it comes to winning games or to farming XP. Also that there is a correlation between win rate and XP, which means WG came with a decent formula for awarding XP.

    The Coward Hindenburg may or may not be playing well, depending on the situation. If he’s alone vs 3 BBs, that’s fine. If he’s at 20km away but could be dealing twice as much damage from 15km, with little risk, then he’s a coward and an idiot (not a star saver, though, as XP comes from damage). Most of the people that annoy teammates with their “star saving cheat” are just that: morons and cowards. They can’t get close enough without getting killed so they stay back and then naturally gravitate towards ships good for the purpose (like Hindenburg). They are also doing the smart thing by picking a low-impact ship if they aren’t all that good – in a Hindenburg or most BBs all they have to do is stay alive, spam damage when possible and let others carry them. Playing this way in a DD won’t work cause that’s against the ship’s role and will be penalized in win rate.

    If you can automate data mining then do this: get a list of all R1’s, get to their personal page and record the stats for ranked T10 games.
    You will get stats of players which have all earned almost the same number of stars (40 stars required from T10 to T1 with only a little noise coming from possible losses when at irrevocable T10. Those are all decently good players so I doubt they spent too much time at the bottom of T10). You can then plot the StarsWon to StarsSaved ratio for players and see if anything interesting happens when compared to XP, average damage or average frags.


  6. “Finally I was accused by mods that I just want to drive traffic to my blog for money, despite the blog doesn’t even have monetization.”

    To be fair, they’re not wrong. Just because your blog isn’t technically monetized doesn’t mean you don’t have a “profit motive” for routing people to it. “Profit” isn’t always measured in money.

    And yeah, confirmation bias is a real thing, deal with it. There is a saying: “You can’t change someone’s mind in real time.” Hammering them with clearly reposted and similar posts like that sure looks like you’re trying to change people’s minds in real time. You can do that on YOUR site no problem, but if you go somewhere else and do that, you might get slapped up a bit.


  7. Oh, those reddit comments are just pure gold.
    Well done Gevlon, well done indeed.

    The question that remains is, how do we best fix or circumvent this effect? How do we show people that they aren’t oppressed but just bad?


  8. @Smokeman: the better question is “would it be bad if my blog was monetized”? Is it somehow a crime to want money for my work? I choose not to monetize because the couple hundred or thousand dollars a year doesn’t worth exactly the above remarks. But those who can’t make it as a work of love and need income are not devils. The doctor who saves your life get paid too.

    There is difference between silently not believing and actively sabotating something you don’t like. And even that is miles away from promoting already falsified fake news. These people were “yes, his dataset is Rickrolling and the text spells BULLSHIT but maybe someone dumber than me misses these”

    @Hanura: that will come soon enough.


  9. There were articles that got accepted to scientific journals and were as fake as that reddit comment.
    I could imagine “standard deviation of correlation” in such a way. Since the data are noisy, correlation can obtain bigger or smaller values. So one can calculate deviations of them.
    In reddit I wanted to see “median absolute difference” near the second “standard deviation”.
    But really amazing way to show that instead of artificial neural networks, we need to study ant algorithm more.


  10. “@Smokeman: the better question is “would it be bad if my blog was monetized”?…”

    No it’s not. That’s an equally irrelevant question. To the end user (Is this context, reddit posters.) it doesn’t really matter WHY you want to route people to your blog, just that you do. “monetization” is just the low hanging fruit of reasons. The mod there doesn’t HAVE to prove, in horrific detail, WHY your posting similar links over and over is actually bad, just that they don’t like it, and any stupid reason is good enough for that.

    You are now engaging in a “Reductio ad absurdum” argument where the ONLY “bad” reason to rout traffic to your blog is monetization.


  11. @Smokeman: since Reddit is a link aggregator, I don’t see why driving traffic to anywhere is bad. The very purpose of Reddit is to drive traffic to interesting things (which is measured by Redditor upvotes)


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 )

Google photo

You are commenting using your Google 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 )

Connecting to %s