1st Of May

1st of May

Miss Atom: Issue 1

zilodak.square.site

Bad Fictional Data vs Fictional Bad Data

WARNING: This post will include discussion of a name that might be Alice Dyer's deadname. I won't be calling Alice by this name or using it in the context of that name being a pointer to Alice, but I will be using the name, uncensored, when talking about where and why the name appeared in chdb.xls .

----

You may know that as part of the ARG, the sleuths on Statement Remains uncovered a document called chdb.xls, allegedly a list that has something to do with The Magnus Institute. There's a list of names, ID numbers, first and last names, dates of birth, and information apparently related to each person's "score" in an assortment of psychological/personality tests. Three of the names in particular have stood out in a lot of analyses: Samama Khalid, Gerard Kaey [sic], and Connor Dyer.

You likely don't know that the commonly linked version of the spreadsheet, ported to Google Docs and linked in the TMAGP ARG Masterdoc, is presented out of order. (I'm guessing they didn't lock down editing until it was already all out of order from various people messing with it - totally understandable, this is not a callout post, thank you for making this easily accessible to people.)

But let me tell you about something I discovered by looking at the spreadsheet in its original order, and the almost certainly incorrect rabbit hole of theorizing it has sent me down.

Bad Fictional Data

Until episode 2 I had the same thought about the Dyer listed in the spreadsheet that I think most people did: that it was Alice's deadname, and that she had therefore been one of the Institute's young subjects. But after Alice had absolutely no reaction when Sam mentioned the Magnus Institute to her in episode 2, I now think this is significantly less likely.

Don't get me wrong: it's still reasonable to think that the Dyer listed in chdb.xls is Alice. Maybe she had some kind of supernatural experience that wiped her memory. (It probably wasn't that Alice was too young to remember, as the Dyer on the spreadsheet is listed as being at Piaget Stage 3, which occurs from 7-11 years old; but it's always possible that the Magnus Institute was using the names of legitimate psychological tests to hide their tracks when recording more esoteric data.) The point is, this isn't hard evidence that Alice has no connection to the Magnus Institute; it just made me go looking for more evidence.

I went back to the spreadsheet to look for more clues about whether or not this was Alice's deadname. What I found instead was some extremely sloppy fake data at the bottom of the spreadsheet.

For context, here are first ten names in the spreadsheet:

A screenshot of 10 rows, 3 columns of a spreadsheet. The columns are labeled I.D., First Name, and Last Name. IDs are in the format of two letters followed by 5 numbers. The two letters of each ID are the first and last initials of the person's name.

Note how each ID begins with the name's first and last initial.

Now check out the last ten names:

10 more rows of the same columns of the spreadsheet. The second letter in the IDs no longer lines up with the last initial of the people's names. Furthermore the IDs display a distinct alphabetical pattern, with CD in the first row, EF in the second, GH in the third, and so on down to UV.

Not only do these IDs no longer always match their subjects' names, they occur in order: CD, EF, GH, IJ, KL, MN, OP, QR, ST, UV. The first names of each pair match the first letter of each ID, but many, though not all, of the surnames don't match the second letter.

My first thought was that whoever Rusty Quill had contracted to generate these names had gotten sloppy at the end of the list, created the IDs all at once using this alphabetical pattern, and picked names to fill in that roughly matched the IDs. But hey, we could use this to our advantage! Any name that was filled in as part of a series of IDs with an alphabetical pattern like this could be removed from consideration for red string analysis - we'd know they were meaningless fakes added by a lazy contractor, and not clues or characters that might show up again later.

Scrolling back up the spreadsheet, we can see the person generating the data having more care the earlier we go. We find the beginning of the AA/BC/DE/FG/HI pattern at line 136, but at first, the names mostly conform to the initials they've been given. JK09874 "Josie Jordan" at line 154 is the first break from the "first two letters of the ID are their initials" pattern; and breaks occur more often the further you go down the sheet.

Scrolling up to before line 136 (AA09911 - Aaron Atkinson), while the pattern isn't yet at AB/CD/EF/GH levels of obviousness, the first initials are still in alphabetical order. Zoe Hart follows Yara Logan follows Xavier Freeman follows Wyatt Edwards. The data creator skips a few letters - for example Niamh Fenton is followed by Phoebe Emmett, and S and T are together in the same line in Skye Travers.

We can follow this less-obvious version of the alphabetical pattern up to an abrupt break right at line 118, above which the IDs don't follow an alphabetical pattern at all. (They might follow a different pattern, but it's not one that I've found yet.) So that means we can discount all the names in line 118 and below as purely fake, generated lazily by a contractor, and not worthy of our attention for the purposes of red-stringing. Right?

A screenshot showing the same spreadsheet but with line numbers showing this time. Line 118 is the first of the lines to follow the sloppy, fake-looking alphabetical order naming scheme. Line 118 has been highlighted in yellow for emphasis. The name in Line 118 is Connor Dyer.

What the fuck?!

(highlight is my own, it is not present in the original document)

My first thought was that the sloppy data generator had done the funniest thing imaginable, sending everyone on a wild goose chase about Alice's deadname just by having the name "Dyer" on the brain while looking for a surname that started with D. This would be Very Funny. No plot relevance, no implications, just the brain fart that launched a thousand theories.

My second thought was that maybe Connor Dyer was the last legit name on the list, and whoever started filling the rest of the sheet in with alphabetical junk data was inspired by the "CD" initials in the first place - whoever it was went on from there.

These are both valid thoughts! But I prefer my third thought:

What if it's on purpose?

Fictional Bad Data

There is a very obvious break between the set of data that doesn't look obviously* fake, and the set of data that is immediately identifiable as such. If we assume that this was intentional - and I want to reiterate that it all being unintentional is still a very real possibility here - why would someone at Rusty Quill want the data to be structured like this?

If the sharp dividing line between reasonable-seeming data and obviously fake data is intentional on RQ's part, it would suggest that we should take the data above row 118 as in-universe real data, and the data below row 118 as in-universe falsified data. It suggests that someone, either at the Institute or after its demise, was adding nonexistent children to the roster of The Magnus Institute. Why would someone want to do that?

There are all kinds of possible reasons, but here are a few off the top of my and my theorizing buddies' heads:

Financial fraud (institutional edition). If the Magnus Institute received funding on a per-child basis, they'd have an incentive to inflate their numbers.

Financial fraud (researcher edition). One or more people on staff were blowing off their child-analysis sessions and recording fake numbers for fake children. This would be ballsy as hell if they could be fired for it, but it was the Magnus Institute, so there's decent odds they couldn't be.

Scientific fraud (faking conclusions edition). The Magnus Institute in the Protocolverse claimed to be doing research on giftedness in children, which is the kind of thing that you'd normally publish in a scientific journal. It's not unheard-of for dickhead academics to falsify data to generate statistically significant results, since statistically insignificant results aren't going to get you published.

Scientific fraud (obscuring paranormal bullshit edition). If the Magnus Institute was using legitimate psychological test names to record Fear-related test results, it's possible their results showed different patterns from what you would expect from the real tests. They could have added the fake children to balance out the dataset as a whole.

Pseudonyms. The children are all real, the Institute just started using fake names for them for privacy purposes. They couldn't go back and change the names they'd already written properly for some reason. Probably something paranormal.

Those are all pretty interesting possibilities, and if we could narrow them down, it might tell us something about what things were like at the Magnus Institute before it burned down!

And the other big question is: why did RQ make the dividing line between the two sections, the first likely-fake entry, Connor Dyer?

One straightforward reason could be as a troll, a red herring to watch fans get in a lather over. And once the community inevitably noticed all the obviously falsified entries, RQ could eat popcorn and watch us lose our minds over whether or not that's even a real entry! (That sounds really fun, I would absolutely do that.)

But let's dig a little deeper, and look at what Connor Dyer being on the border between the real and fake entries would mean in-universe. Because of its position as the border between real and fake, it would be very easy for that entry to be accidentally included in the wrong group - a real research subject discarded as fake, or perhaps more interestingly, a fake research subject accidentally reclassified as real.

Remember, if a name is fake in the context of the Magnus Institute's research, that doesn't mean that the name itself is made up. If I was trying to think of a name that fit the initials CD, and those were the initials of my next door neighbor's kid, I might just write their name in as a lark. Especially if it was my first time trying to get away with falsifying information: this is a kid that verifiably exists and lives in the area.

My theory, supported primarily by my love for The Implications instead of actual evidence

Twenty years on, after all institutional memory of the fraud was long gone, trans icon Alice Dyer applies to work for the OIAR - an institution that (according to this theory) has an unofficial preference for hiring former Magnus Institute kids.

They are very confused when Alice proceeds to act nothing like a former Magnus Institute kid. It doesn't occur to anyone that her entry might have been falsified. What reason would anyone have to do that?

----

* Of course people with a background in data analysis or statistics will see immediately that even above line 118 this is a wild-ass dataset that would raise red flags for falsification, but at least it's not "the alphabet over and over" levels of obvious.

174 notes View Post

rookrock

10 months ago

Wilding out Celia theory:

- There's two Celias

- One from this universe, one from TMA

- They share a body

- When Celia wakes up after sleepwalking, that's the Celia of TMAGP

- When Celia's at the OIAR, that's the Celia of TMA

- one of them"s a baddie

*tinfoil noises*

- TMA Celia has become an avatar of the stranger