How To Lie With Statistics

2021-04-07

Header art by Cryo Huren

Has the Imperium lost all its sovereignty and is now on the verge of losing the war?  Are TAPI and BRAVE bleeding the other members of Legacy dry, leaving them to be dumped on the ash heap of history?  Are titan prices going up by 200%?  Is EVE dying?

Have you ever wondered how the people who discuss topics seem to have the data and numbers that support their position?  True, sometimes the data does support their argument.  Other times, it can be tempting to manipulate numbers and figures to match the agenda you are presenting.  People tend to gravitate toward the statistics which match their agenda, even if they are not consciously doing it.  For example, advertisers, companies, and politicians will present you with what seems like hard believable numbers and charts.  However, statistics are not as objective as they seem.  People don’t even need to lie; they just need to flex the numbers.  They change the dimensions and lengths of time a chart tracks.  They present a series of data sets but leave out correlating data that would give it more context, or would change the picture being presented.

Charts like this one can be used, improperly, to indicate a failscade in Legacy.

I will pick on The Imperium first, since I am a Goon.  The narrative that has been presented recently is that TEST is evil.  They treat BRAVE like crap.  They’re leaching all the corporations, money, and sov from the other Legacy members.  They’re leading them out of their traditional homelands to their doom.  Any free-thinking, intelligent member of Legacy should leave the coalition for their own good; correct?

If you pull up Dotlan maps, zkillboard, and follow Imperium News the numbers and charts presented would seem pretty solid.  If you run the numbers, they actually support the narrative you have just been presented.  Rezeda Regnum (R-R) has lost half their pilots, half their corporations, and all their sov space since the war began.  Even if TEST claims Legacy is winning the war, obviously R-R isn’t.  The Watchmen have ceased to exist.  Evictus has lost about 15% of their members, corporations, and sov.  Warped Intentions has lost 28% of their pilots, corporations, and almost all of their sov.  Federation Uprising (FEDUP) has lost most of their established sov, and although they’ve taken new systems, they have to rebuild all that infrastructure.  Eternal Requiem (REQ.) is down 15% and lost all its sov.  VINDICTIVE (VINDI) numbers are down 15%.  On paper, it seems catastrophic.  Someone not familiar with EVE Online, when presented with these “facts” might almost certainly believe that Legacy was starting to lose the war.  But let us examine these numbers from a different angle.

Legacy Coalition right now has around 12 Alliances in it and 45,000 pilots serving in it.  I’m not counting Legacy Affiliates, PanFam, PanFam Affiliates, Winter Coalition, or any of the proxy wars going on.  Out of those 45,000 pilots most of them come from TEST and BRAVE.   These Legacy members have gained strength during the war, even if we look at BRAVE’s structure and sov losses.  That means 80% of Legacy’s fighting strength comes from those alliances that are actually stronger than they used to be, based on number count.  Therefore, the statistics in the paragraph I presented regarding the other nine alliances now has significantly less impact.  Depending on what numbers and charts I show, what order they’re presented in, and other data sets I decide to put in or leave out, I can paint an entirely different picture of what’s happening in the war.  Depending on the readers’ allegiance, narrative, preconceived ideas of what’s going on with the war, this report now has significantly more, or less, impact on them depending on the effect I want to achieve. I can lie with statistics.

Now I get to pick on Legacy.  The narrative has been presented that Goonswarm Federation is evil.  We’re crushing all the life and fun out of the game, and we need to cease to exist as a coalition for the good of all.  If you look at Sov.space maps and check sovereignty settings on Dotlan, the argument could be made that we have been crushed.  Most of the former Imperium territories in Delve now belong to TEST.  PAPI has swallowed more than half the map.  The Imperium is done for.  The giant “Mission Accomplished” banner is hanging on an Avatar while Vily makes speeches.  So, let’s examine these numbers and “facts” from a different angle.

Remember those Legacy pilot numbers from earlier?  45,000 pilots serving in Legacy.  Probably at lease 25k to 30k pilots deployed to Delve as the main front of the war.  Goonswarm Federation currently has over 29,000 pilots, 500 corporations, and still holds 17 Sov systems (at the time of this writing) even though we don’t currently appear on the map.  The Imperium has over 40,000 pilots, a big chunk of whom are concentrated in Delve around the Imperium Capital system of 1DQ.  After nine months at war, active pilot numbers have not gone down.  Most of the Imperium’s Capital Ship Fleet is intact and deployed to fight in the coming battles.  Goonswarm Federation is very much alive.

All the statistics you will ever hear are based on samples.  Inevitably, these samples have a built in bias.  Despite people’s best efforts and intensions, everyone is biased in some way.  When you are looking into something you are really interested in, about which you really want to learn the truth, you need to consider the motives of those who are providing you with information and the accuracy of their findings.  Someone once said knowledge is a three-edged sword.  Your side, their side, and the truth.  In the universe of EVE Online, the truth can be a fickle thing, especially when statistics are involved.

Let your voice be heard! Submit your own article to Imperium News here!

Would you like to join the Imperium News staff? Find out how!

Comments

  • Marus

    Thumbs up for an article that comes the closest to the reality of the events and the truth.

    April 7, 2021 at 8:49 AM
  • Gwailar

    The Imperium narrative which you assess you state thus:

    “The narrative that has been presented recently is that TEST is evil. They treat BRAVE like crap. They’re leaching all the corporations, money, and sov from the other Legacy members. They’re leading them out of their traditional homelands to their doom.”

    After presenting statistics which show strong declines in Legacy’s minor powers but an overall steady member count for the coalition, you assert your counter narrative thus:

    “That means 80% of Legacy’s fighting strength comes from those alliances that are actually stronger than they used to be, based on number count. Therefore, the statistics in the paragraph I presented regarding the other nine alliances now has significantly less impact.”

    In fact what you’ve done is prove the Imperium narrative you seek to deconstruct. That narrative–even as you state it–isn’t that Legacy is weaker over all. It’s that while TEST flies high, the rest of Legacy slowly bleeds. That TEST has lead the rest of Legacy into a war that’s great for TEST and terrible for almost everyone else. That narrative is most certainly correct, even by your own analysis.

    You need to write your next article on “lying by changing the terms of the argument mid stream.”

    April 7, 2021 at 11:58 AM
    • Michael Porter Gwailar

      The point of the article is to show you can use data to make any argument/narrative you want, not to deconstruct the Imperium’s views. It’s true, and it’s not entirely true, but nothing is. Also, huge respect for you writing. Sincerely.

      April 7, 2021 at 12:11 PM
  • Rammel Kas

    The one reason I find people reaching for statistics as often as they do is that they’re looking for indicators based on their experience. And there sure are plenty of those about, as you’ve alluded to. Since nobody can accurately get inside the head of other players much less in groups larger than Dunbar’s number watching some measurable results of what they do in the game is perhaps the next best thing.

    April 7, 2021 at 11:59 AM
  • Winston Archer

    Two thumbs up. Succinct, informative, and no spin. Thank you sir. o7

    April 7, 2021 at 12:05 PM
  • Havish Montak

    I have decided that this article is 100.1% spin.

    April 7, 2021 at 1:15 PM
    • You are right, if it is more than 100% spin then it means it is in the direction of truth. Thx for confirmation

      April 8, 2021 at 8:38 AM
  • Deni'z von Meanace

    Whatbis that about? Hey look Legacy is not so strong as before, look at PAPI’s proxy, they are dying. But Imp is strong, it doesnt matter its trapped arround 1DQ and don’t fight anymore and didn’t win any major battle for the last month. But you keep saying Imo is strong, strong in what? Propaganda articles? Init and Co do more actions at TEST land vs Goons at Delve. So now its obvious vice versa where Goons don’t respond to a fight and those days where they blame test being cowards are over. Reading all that articles in the wars historical perspectives, bringing examples from the past don’t chnage the overall position of Goons. You guys done.

    April 7, 2021 at 11:37 PM
    • Try this again in English, please?

      April 8, 2021 at 2:02 AM
      • Moomin Amatin Arrendis

        Allow me,
        GRR GON
        HAT GON
        GON DED

        April 8, 2021 at 3:14 AM
    • Bumpy Dog Deni'z von Meanace

      Incoherent word salad.

      April 8, 2021 at 7:16 AM
  • Bronze Condor

    Should Test be opposed?

    April 8, 2021 at 12:41 AM
  • BriscRubal

    This is a really bad take, for a variety of reasons. First, the statistics are the statistics. Legacy is more than just Brave and TEST. Those groups in their areas held a lot of Sov, and they were and are expected to be part of the war effort. That their numbers are being chopped down, they’re losing sov, and many of them are leaving the war entirely is a legitimate narrative that’s demonstrated by the data. To pretend that Legacy is fine because Test is fine (Brave certainly isn’t) is bad analysis that’s not supported by the facts. And there are more facts than slopes and sov – activity levels and the rest also demonstrate that these groups are indeed failscading and they are no longer as active as they were.

    One of the reasons smaller groups join coalitions like this is to get the benefit of having larger friends who will help out and protect them. That’s the only reason, for instance, Brave is even in Legacy and that comes directly from Dunk Dinkle. When you see smaller groups in a coalition fall apart because they aren’t getting support from the big guys, and somebody calls that fact out, that’s not data being used improperly. That’s data being used to justify an argument.

    If you really disagree with those who are arguing based on the data that Legacy is in trouble, you’ll have to do more than rip off Mark Twain and act like a slope isn’t really a slope because *reasons*.

    April 8, 2021 at 12:50 AM
    • Moomin Amatin BriscRubal

      You should have got the the last paragraph I think as that is where the “meat” is. This is an Eve take on “lies, damned lies and statistics” I think. The “details” to support that mean nothing other to bolster the argument that truth is often subjective.

      April 8, 2021 at 1:52 AM
      • BriscRubal Moomin Amatin

        That’s the problem. All we can do is make an argument, and invite others to agree or disagree. There is no objective truth in a situation like this, where you’re dealing with a ton of outside variables and human irrationality thrown into the mix.

        We can look at the data and tell you what is happening. We can’t always tell you with any precision why. That’s part of the reason why this article annoys me, beyond that it’s basically saying the Meta Show is full of shit, which we aren’t.

        April 8, 2021 at 2:23 AM
        • Moomin Amatin BriscRubal

          I feel your pain. This is why I went back and looked at the member count according to Dotlan. It is flawed data as has been covered but it is at least consistent. When I did look at the data it turns out that of the three blocs that have formed, FIRE + Legacy – PanDaFam – The Imperium, it was The Imperium who were smaller in member count and space owned. I love this war and the optics in the future are going to be amazing.

          April 8, 2021 at 3:09 AM
    • Marus BriscRubal

      If I were you I would stick at commenting my own coalition status. The imperium is not just goonswarm and init. Why don’t you tell us how bastion and tnt are doing? Or even init for that matter. Sure goonswarm remains strong but can you say the same for init, bastion and tnt? Aren’t they putting these smaller aliances in the same position you say test puts those in legacy?
      Like I said before, your shit smells just as bad as the rest of eve’s alliances, you just don’t want to admit it and try to hide it with a cheap parfume wich makes it actually worse.

      April 8, 2021 at 8:50 AM
      • BriscRubal Marus

        Yes, I can say the same for INIT, Bastion and TNT. Bastion is wrecking Esoteria and has been single handedly. INIT lost Fountain, regained Fountain, took Impass and Catch, and is losing Fountain again. Our activity numbers are still pretty high – where our numbers look down is we’ve not been using our titans and supers. TNT is active in Delve. The Imperium has lost no major groups, unlike Legacy.

        April 8, 2021 at 1:30 PM
        • Seir Luciel BriscRubal

          I think you would at least concede Adreland’s point, Brisc. I get that you don’t buy his examples and that’s fair. But the guy’s point is that we all ought use critical thinking when we encounter statistics and ask good questions, in EVE, as everywhere else. Last I knew you worked in politics; I know you know the shenanigans of statisticians as much as anyone.

          April 8, 2021 at 2:13 PM
          • BriscRubal Seir Luciel

            When you are building arguments in any kind of advocacy situation, you need facts to back up your conclusions and why you’re making the argument in the first place. One of the obvious and most common ways to do that is with statistics. The whole “lies, damned lies and statistics” thing is funny, and it’s true that depending on how you frame them you can make an argument look better or worse, but in the end, this is actual data.

            Data, alone, is meaningless. It’s just numbers. You need to analyze it and explain what it means for it to have value.

            April 9, 2021 at 1:27 PM
          • Seir Luciel BriscRubal

            You’re right: data, alone, is meaningless. “That’s the “statistics don’t lie” half of the saying.

            The difference between “lying” and “just a bad argument” is an interesting one. I might even say it is an ideological one; one could argue that lies, by definition, (among other things) are also bad arguments. What distinguishes the two as we define them: The intention of the speaker? What knowledge the speaker has vs chooses to reveal when addressing an audience? Sincerity?

            I don’t actually think we disagree with each other. I think we are getting tangled in semantics.

            April 10, 2021 at 3:24 PM
      • Rhivre Marus

        Hi there. Tone down the vitriol in your comments.

        April 8, 2021 at 2:33 PM
      • Garreth Vlox Marus

        “Sure goonswarm remains strong but can you say the same for init, bastion and tnt”

        You mean the guys who have spent the last few months running a bulldozer through test’s home regions to hilarious effect totaling trillions in damage to TAPI? Yeah their doing fine, or don’t you know how to read Zkill?

        April 14, 2021 at 2:18 AM
    • Michael Porter BriscRubal

      Again, the article was just an analysis of statistics in general. Different ways of looking at things. The title is based on the book “How To Lie With Statistics” by Darrell Huff. It was not a stab at Gwailer’s article. I wasn’t snubbing the Meta Show. Legacy can get f&*%ed, I don’t have an agenda. I build ships for the Imperium, and write. Huge respect Brisc.

      April 8, 2021 at 3:26 PM
      • BriscRubal Michael Porter

        Again, I appreciate that, but we’ve done two shows where we’ve gone over in detail the “statistics” that you pooh-poohed here and drew the conclusion you said was not fair to draw. I think you’re wrong, and I take umbrage over the claim that what we were doing was lying with statistics. We made an argument, and explained, using the data we had available, what we saw happening. You can agree or disagree, but the data is the data.

        April 9, 2021 at 1:32 PM
        • Garreth Vlox BriscRubal

          I fucking love watching someone use the word umbrage correctly.

          April 14, 2021 at 2:15 AM
  • Susurrus Synaesthesia

    “There are three kinds of lies: lies, damned lies, and statistics.”
    – Unknown/disputed

    April 8, 2021 at 9:14 PM