#That isn't right

[cacw]

sometimes i remember davekat is canon and i get scared

[themistymountainscold]

what do you mean crowley had brown hair season 1

[clarrisani]

Seems to be true of 2024 as well.

AO3 Ship Stats: Year In Bad Data

You may have seen this AO3 Year In Review.

It hasn’t crossed my tumblr dash but it sure is circulating on twitter with 3.5M views, 10K likes, 17K retweets and counting. Normally this would be great! I love data and charts and comparisons!

Except this data is GARBAGE and belongs in the TRASH.

I first noticed something fishy when I realized that Steve/Bucky – the 5th largest ship on AO3 by total fic count – wasn’t on this Top 100 list anywhere. I know Marvel’s popularity has fallen in recent years, but not that much. Especially considering some of the other ships that made it on the list. You mean to tell me a femslash HP ship (Mary MacDonald/Lily Potter) in which one half of the pairing was so minor I had to look up her name because she was only mentioned once in a single flashback scene beat fandom juggernaut Stucky? I call bullshit.

Now obviously jumping to conclusions based on gut instinct alone is horrible practice... but it is a good place to start. So let’s look at the actual numbers and discover why this entire dataset sits on a throne of lies.

Here are the results of filtering the Steve/Bucky tag for all works created between Jan 1, 2023 and Dec 31, 2023:

Not only would that place Steve/Bucky at #23 on this list, if the other counts are correct (hint: they're not), it’s also well above the 1520-new-work cutoff of the #100 spot. So how the fuck is it not on the list? Let’s check out the author’s FAQ to see if there’s some important factor we’re missing.

The first thing you’ll probably notice in the FAQ is that the data is being scraped from publicly available works. That means anything privated and only accessible to logged-in users isn’t counted. This is Sin #1. Already the data is inaccurate because we’re not actually counting all of the published fics, but the bots needed to do data collection on this scale can't easily scrape privated fics so I kinda get it. We’ll roll with this for now and see if it at least makes the numbers make more sense:

Nope. Logging out only reduced the total by a couple hundred. Even if one were to choose the most restrictive possible definition of "new works" and filter out all crossovers and incomplete fics, Steve/Bucky would still have a yearly total of 2,305. Yet the list claims their total is somewhere below 1,500? What the fuck is going on here?

Let’s look at another ship for comparison. This time one that’s very recent and popular enough to make it on the list so we have an actual reference value for comparison: Nick/Charlie (Heartstopper). According to the list, this ship sits at #34 this year with a total of 2630 new works. But what’s AO3 say?

Off by a hundred or so but the values are much closer at least!

If we dig further into the FAQ though we discover Sin #2 (and the most egregious): the counting method. The yearly fic counts are NOT determined by filtering for a certain time period, they’re determined by simply taking a snapshot of the total number of fics in a ship tag at the end of the year and subtracting the previous end-of-year total. For example, if you check a ship tag on Jan 1, 2023 and it has 10,000 fics and check it again on Jan 1, 2024 and it now has 12,000 fics, the difference (2,000) would be the number of "new works" on this chart.

At first glance this subtraction method might seem like a perfectly valid way to count fics, and it’s certainly the easiest way, but it can and did have major consequences to the point of making the entire dataset functionally meaningless. Why? If any older works are deleted or privated, every single one of those will be subtracted from the current year fic count. And to make the problem even worse, beginning at the end of last year there was a big scare about AI scraping fics from AO3, which caused hundreds, if not thousands, of users to lock down their fics or delete them.

The magnitude of this fuck up may not be immediately obvious so let’s look at an example to see how this works in practice.

Say we have two ships. Ship A is more than a decade old with a large fanbase. Ship B is only a couple years old but gaining traction. On Jan 1, 2023, Ship A had a catalog of 50,000 fics and ship B had 5,000. Both ships have 3,000 new works published in 2023. However, 4% of the older works in each fandom were either privated or deleted during that same time (this percentage is was just chosen to make the math easy but it’s close to reality).

Ship A: 50,000 x 4% = 2,000 removed works Ship B: 5,000 x 4% = 200 removed works

Ship A: 3,000 - 2,000 = 1,000 "new" works Ship B: 3,000 - 200 = 2,800 "new" works

This gives Ship A a net gain of 1,000 and Ship B a net gain of 2,800 despite both fandoms producing the exact same number of new works that year. And neither one of these reported counts are the actual new works count (3,000). THIS explains the drastic difference in ranking between a ship like Steve/Bucky and Nick/Charlie.

How is this a useful measure of anything? You can't draw any conclusions about the current size and popularity of a fandom based on this data.

With this system, not only is the reported "new works" count incorrect, the older, larger fandom will always be punished and it’s count disproportionately reduced simply for the sin of being an older, larger fandom. This example doesn’t even take into account that people are going to be way more likely to delete an old fic they're no longer proud of in a fandom they no longer care about than a fic that was just written, so the deletion percentage for the older fandom should theoretically be even larger in comparison.

And if that wasn't bad enough, the author of this "study" KNEW the data was tainted and chose to present it as meaningful anyway. You will only find this if you click through to the FAQ and read about the author’s methodology, something 99.99% of people will NOT do (and even those who do may not understand the true significance of this problem):

The author may try to argue their post states that the tags "which had the greatest gain in total public fanworks” are shown on the chart, which makes it not a lie, but a error on the viewer’s part in not interpreting their data correctly. This is bullshit. Their chart CLEARLY titles the fic count column “New Works” which it explicitly is NOT, by their own admission! It should be titled “Net Gain in Works” or something similar.

Even if it were correctly titled though, the general public would not understand the difference, would interpret the numbers as new works anyway (because net gain is functionally meaningless as we've just discovered), and would base conclusions on their incorrect assumptions. There’s no getting around that… other than doing the counts correctly in the first place. This would be a much larger task but I strongly believe you shouldn’t take on a project like this if you can’t do it right.

To sum up, just because someone put a lot of work into gathering data and making a nice color-coded chart, doesn’t mean the data is GOOD or VALUABLE.

[bandzboy]

since spotify wrapped is coming today i just want to bring awareness to the fact that around 86% of the songs on spotify are currently demonetized since spotify decided that tracks with under 1,000 cannot be monetized. not only that the ceo of spotify cashed out $35.8 million dollars in spotify shares in the third quarter of the year and it has been reported that he earns more than the top artists on the platform.

for a platform that claims to support artists, this is outrageous and i hope people realize that an artist who is starting out, cannot make a living out of spotify streams simply because daniel ek and his friends made it worse for the artists trying to start a career. if you wanna support musicians and the possibility for them to get a living wage please follow United Musicians and Allied Workers (UMAW) which are trying to make The Living Wage for Musicians Act a reality so musicians can be paid fairly through streaming platforms and get the cut they deserve.

[poorly-drawn-mdzs]

Expertise can't help you here.

[lazylittledragon]

can't believe we're all adults being forced into the club penguin level of censorship in 2024

[donutdrawsthings]

People all over the world are thinking of you!

[bereft-of-frogs]

There’s that post that’s like ‘everyone should get into a tiny niche fandom at least once’ fully agree, that was really fun -- but I would like to add that everyone should get into a fandom where their opinions run counter to major fanon because it really teaches you about sticking to your guns and trusting your interpretation of the text without having to rely on peer validation

because WHAT are people talking about sometimes

[wiisagi-maiingan]

People who are incarcerated deserve the right to vote. I cannot believe that the biggest argument against that is still "well they'd vote to legalize [murder/rape/etc]" as if ballots have a fill-in-the-blank option for new laws. So many aspects of our society are left up to people who have absolutely no idea how any of it works.

[egophiliac]

looking at next month's schedule and between the end of 7-12 and the wishing lantern event it's like

February is officially RIDDLE MONTH, brace yourselves to be absolutely blasted into ashes everybody

[bulionis]

stobotnik kissy I drew for reaching 500 followers on Twitter 😳

[fascinationstreetmp3]

annoying4annoying (insp)

[adobe-outdesign]

Babe wake up, massive GameFreak leak just dropped

Literally like a terabyte of info just came out and stuff's still leaking in mass from Twitter as I type so bear with me. I'm only including a few highlights (including the most legit-seeming ones) because there's potentially over a hundred beta sprites/art/assets being leaked.

All of these leaks were pulled from r/PokeLeaks. I'm doing my best to verify as I go, but will update the post if any of this turns out to be fake (though this is unlikely).

^ Unknown water starter concept art

^ Unused Pokemon

^ Unused Pokemon part 2 electric boogaloo

^ Beta Honchkrow with a mustache (that explains the random crescent around its eye in the final version)

^ fear.jpg (Beta Kricketune)

^ Beta wash Rotom

^ Beta Lickilicky (now we know why it evolves via rollout in the final version)

^ Beta Skunktank

^ Beta Reshiram/Zekrom

^ Beta team Galactic admins

[nyancrimew]

last thing i'll say about US politics on here is that you'll never get a more left wing president if you just shut up and tell everyone around you to shut up too and leave criticisms of democrats to the right. criticizing a candidate from the left doesn't just magically convince people to vote for the right wing candidate instead and is the only way for the establishment to actually see what their voter base wants. you will never get change or progress from just shutting up and letting things happen because the alternative requires effort and seems scary.

when biden suddenly announced all of these great things for his second term why do you think this happened? because you kept posting about how we just have to vote for him and stop debating stuff or because of the people who demanded more from him and made him question how much support he actually has?

[poorly-drawn-mdzs]

You're just not toxic enough.

[biggest-gaudiest-patronuses]

genuinely is my brain erasing itself or is google/tumblr censoring that post about notre dame and old growth forests.