![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
letzanHigh-level stats for week of 2021-03-23 - 2021-03-29
- Total works categorized F/F on AO3: 6974 (-67 from last week)
- Works I classified F/F: 4304 (-53 from last week) (1932 new, 2372 continued)
- 1.03% of all 418882 AO3 works I've classified F/F were updated this week
| Before algorithm change | After algorithm change |
|---|---|
 |
 |
 |
 |
A few callouts this week:
- Hi readers. So, I rearranged the deck chairs a bit this week, in a way that is going to make the rankings subtly but significantly different. If anyone pays attention to what fandoms show up in the top 20 week after week, you'll want to give this a read.
- What changed? As you may know, my ratings in the top 20 are a couple of steps removed from being a pure count of works tagged F/F. First, I include only works which are actually tagged with at least one ship containing multiple canonically non-male fictional characters. Second, I apply a "tag-scroller frustration" discount rate, so that fandoms where many works which are labelled F/F don't actually contain a femslash ship, get a worse rating than works where F/F-labelled works reliably contain femslash ships. Third, because I am personally a fan of ensemble fandoms with a lot of characters and ships, I use a logarithm of the number of works per ship in a fandom, so that a fandom with one giant ship won't do as well as a fandom with two ships half the size. Or I did. The change is that I'm no longer doing that third thing. Now the rating just counts the number of works updated this week tagged with the fandom and categorized F/F which contain a femslash ship, and applies the anti-frustration discount rate.
- That was complicated. Why the change? Basically I decided to make the change as a result of trying to compare Destination Toast's analysis to my own a few weeks ago. I kept having to put caveats into my analysis because I have this weird factor that changes the ratings of fandoms with similar work counts in hard-to-think-about ways. And I realised that it's often the case that I have a hard time explaining in my own comments why a particular fandom with a low work count is where it is on the charts, or why it's on the charts at all, and that's not helping anybody understand the state of femslash better. Don't get me wrong; I still like ensemble fandoms. But it's more important to have a rating system that we can have conversations about, than it is to account for an arbitrary feature of my personal fandom taste. So it's time to rip that out and compute the stats in a simpler way.
- Does the new algorithm change anything? Yeah, it totally does, which is why I wrote so many words about this. I don't know how the formatting on this post will look, but I'm going to at least try to include before and after versions of some graphs that will show the big differences. Most obvious is that Supergirl (a fandom with an extremely dominant top ship, especially these days) is no longer having any trouble staying at number one. Previous challenger Harry Potter was using ship diversity to make up for low F/F fidelity (only 60% of Harry Potter works tagged F/F in a given week actually contain a ship with multiple canonically non-male characters, versus nearly 100% for Supergirl), and RWBY was using ship diversity to make up for not having nearly as many works per week as Supergirl.
- What else will we see more of from now on? Several other dominant-ship fandoms benefitted from the change: She-Ra, Owl House, The 100, Once Upon a Time, Wynonna Earp, The Wilds. Most of those fandom names aren't new to readers of the top 20, but they're doing better now, and that'll probably keep happening. Most dramatically, it's worth flagging Attack on Titan, which has been seen in the top 20 but not since 2017, in the same sentence with Haunting of Bly Manor, which has only one represented ship and was never in the top 20 prior to this change, but which has 50+ works a week. Neither of those fandoms is in the top 10 even now --- I did the math on this awhile ago and I was pretty sure I couldn't be missing a fandom with reliably top 10 work count, but they properly should have been in the top 20 and now they are.
- What will we see less of? Well, the obvious answer is anything that doesn't have a dominant ship. So MCU is going to fall in the ranks (especially because it has terrible fidelity, e.g. 52% this week), as is Avatar: the Last Airbender, as is Dangan Ronpa, as is Fire Emblem. Most of those fandoms aren't going to disappear entirely, but they'll definitely spend less time in the top 5 or top 10, or even on the charts at all. We might lose fandoms like BanG Dream! or Love Live! entirely, but then again we might not --- those fandoms have had competitive work counts at some times in the past. Anyway, this is all going to be interesting, and thanks for bearing with me as I quietly change how everything works.
- If anyone is still reading after all that, I'll reward you by talking about an actual fandom thing: while looking over the table, I realised that the top pairing in Harry Potter is suddenly no longer Ginny/Luna or Hermione/Bella or Pansy/anyone, but Dorcas Meadowes/Marlene McKinnon. This isn't precisely a new pairing --- there's works for it as far back as 2010 --- but there's been a notable ramp-up in work count since December or so, and it's had 30+ works/week for the past two months. Paging through recent works, my guess is that it's a common secondary pairing in marauders fic, and that's just been through the roof overall recently. But I do often like wartime and post-war fic, so have a couple of recs like that: here, here, and here.
Most active classified fandoms:
| Rank | Fandom (and top three ships) | Rating for week | F/F works updated this week | % of F/F works updated this week | F/F:M/M ratio (all works) |
|---|---|---|---|---|---|
| 1 | Supergirl (TV 2015) Kara Danvers,Lena Luthor (151) Alex Danvers,Samantha "Sam" Arias (21) Alex Danvers,Maggie Sawyer (20) | 195.04 | 198 (+21) | 0.79% | 28:1 |
| 2 | She-Ra: Princess Of Power Adora (She-Ra),Catra (She-Ra) (129) Perfuma (She-Ra),Scorpia (She-Ra) (27) Catra (She-Ra),Glimmer (She-Ra) (11) Netossa (She-Ra),Spinnerella (She-Ra) (11) | 136.00 | 136 (-25) | 1.77% | 5:1 |
| 3 (↑2) | RWBY Blake Belladonna,Yang Xiao Long (83) Penny Polendina,Ruby Rose (RWBY) (19) Blake Belladonna,Weiss Schnee (11) | 132.94 | 149 (+15) | 1.43% | 3:1 |
| 4 | The Owl House (Cartoon) Amity Blight,Luz Noceda (112) Emira Blight,Viney (The Owl House) (24) Boscha (The Owl House),Willow Park (14) | 128.02 | 130 (-4) | 4.66% | 16:1 |
| 5 (↓2) | Harry Potter - J. K. Rowling Dorcas Meadowes,Marlene McKinnon (52) Ginny Weasley,Luna Lovegood (30) Hermione Granger,Pansy Parkinson (18) | 119.88 | 201 (-33) | 1.81% | 1:7 |
| 6 | The 100 (TV) Clarke Griffin,Lexa (The 100) (86) Anya (The 100),Raven Reyes (19) Costia (The 100),Lexa (The 100) (9) Luna (The 100),Raven Reyes (9) | 95.02 | 97 (+13) | 0.70% | 5:1 |
| 7 (↑2) | Legacies (TV 2018) Hope Mikaelson,Josie Saltzman (51) Josie Saltzman,Penelope Park (27) Hope Mikaelson,Lizzie Saltzman (17) | 87.00 | 87 (+11) | 3.05% | 13:1 |
| 8 (↑2) | Once Upon a Time (TV) Emma Swan,Evil Queen | Regina Mills (75) Belle (Once Upon a Time),Red Riding Hood | Ruby (7) | 84.05 | 86 (+13) | 0.58% | 9:1 |
| 9 (↑2) | Dangan Ronpa - All Media Types Mioda Ibuki,Tsumiki Mikan (17) Chabashira Tenko,Yumeno Himiko (15) Akamatsu Kaede,Iruma Miu (13) | 80.00 | 100 (+25) | 2.38% | 1:3 |
| 10 (↓3) | 僕のヒーローアカデミア | Boku no Hero Academia | My Hero Academia Jirou Kyouka,Yaoyorozu Momo (78) Asui Tsuyu,Uraraka Ochako (33) Toga Himiko,Uraraka Ochako (13) | 68.75 | 137 (-15) | 2.09% | 1:8 |
| 11 (↑) | The Wilds (TV 2020) Shelby Goodkind,Toni Shalifoe (45) Fatin Jadmani,Leah Rilke (33) | 67.00 | 67 (+8) | 11.38% | 332:1 |
| 12 (↑6) | Avatar: The Last Airbender Kya II (Avatar),Lin Beifong (19) Azula (Avatar),Ty Lee (Avatar) (15) Mai (Avatar),Ty Lee (Avatar) (11) | 66.61 | 73 (-2) | 2.36% | 1:1.7 |
| 13 (↑2) | 原神 | Genshin Impact (Video Game) Jean (Genshin Impact),Lisa (Genshin Impact) (21) Beidou (Genshin Impact),Ningguang (Genshin Impact) (20) Ganyu (Genshin Impact),Keqing (Genshin Impact) (11) | 59.84 | 75 (+0) | 9.20% | 1:6 |
| 14 (↑2) | Wynonna Earp (TV) Nicole Haught,Waverly Earp (51) Waverly Earp,Wynonna Earp (15) Nicole Haught,Wynonna Earp (10) | 52.00 | 55 (+1) | 1.08% | 18:1 |
| 15 (↓7) | Shingeki no Kyojin | Attack on Titan Krista Lenz | Historia Reiss,Ymir (Shingeki no Kyojin) (68) Annie Leonhart,Mikasa Ackerman (21) Mikasa Ackerman,Sasha Blouse (5) | 50.61 | 88 (-26) | 2.20% | 1:6 |
| 16 (↑3) | Marvel Cinematic Universe Natasha Romanov (Marvel),Wanda Maximoff (20) Maria Hill,Natasha Romanov (Marvel) (17) Agatha Harkness,Wanda Maximoff (10) | 49.74 | 100 (+30) | 0.96% | 1:8 |
| 17 (⇑) | Fire Emblem: Fuukasetsugetsu | Fire Emblem: Three Houses Dorothea Arnault,Ingrid Brandl Galatea (8) Dorothea Arnault,Edelgard von Hresvelg (7) Hilda Valentine Goneril,Marianne von Edmund (7) | 48.09 | 69 (?) | 2.56% | 1:2 |
| 18 (↓) | The Haunting of Bly Manor (TV) Dani Clayton,Jamie (The Haunting of Bly Manor) (42) | 42.00 | 42 (-8) | 4.50% | 529:1 |
| 19 (⇑) | Doctor Who Thirteenth Doctor,Yasmin Khan (31) River Song,Thirteenth Doctor (11) Clara Oswin Oswald,Thirteenth Doctor (3) | 41.14 | 48 (?) | 0.88% | 1:1.1 |
| 20 (⇑) | Miraculous Ladybug Juleka Couffaine,Rose Lavillant (20) Chloé Bourgeois,Marinette Dupain-Cheng (14) Alya Césaire,Marinette Dupain-Cheng (11) | 40.31 | 54 (?) | 2.35% | 1.1:1 |
What is all this again?
Mission statement: help out people (including myself) who read fic on AO3 and wish they were reading more femslash, but don't know what's out there. Plan: do a week-in-review post every week which contains general stats about new fic and a few recs.
I am starting with the list of all works on ao3.org which were updated within a given week, and which were categorized as F/F, but I am not reporting stats on 100% of those works --- I'm narrowing down to a subset of fandoms and works for which it's tractable for me to actually look at all the character tags, and then I'm excluding ships containing what I believe to be cis-male characters. I'll give overall stats about how many works I've reported on and how many I've excluded. If you think I'm excluding a fandom you care about, please let me know!
More info about column headers:
- Fandom (and top three ships): If multiple tags that seem to correspond to the same fandom (e.g. "The 100 (TV)" and "The 100 Series - Kass Morgan") show up in my ranking, I report the one with the best rating. There's totally a judgment call here about what things are "the same fandom" --- I'm going to err on the side of reporting more different things when I can, but I won't skip something with a lot of works which aren't otherwise accounted for. When measuring the top three F/F ships in a fandom, I count both "A/B" and "A & B" as "A,B". Ties are broken by including all of the ships, but I only include a maximum of five ships per fandom to keep the table length reasonable.
- Rating for week: This is the number of works I classified as F/F, multiplied by what I call "F/F fidelity", which is the fraction of all works tagged F/F on AO3 that I classified as containing at least one F/F ship.
- F/F/ works updated this week: includes both new and continued works.
- % of F/F works updated this week: This is the percentage of all F/F works I've ever categorized in the fandom, so a big number probably means a brand new fandom, or at least one which was unusually active this week.
- F/F:M/M ratio (all works): This ratio is taken from all works in the fandom tag, and is just a hint to readers about how F/F-dominant the fandom is. Obviously there's no reason you can't find great F/F works in a predominantly male fandom (we've all read those), but the ratio may give a hint about whether the F/F pairing is likely to be the dominant one in a work with multiple pairings. Note that the "F/F" and "M/M" numbers for this column are taken directly from AO3, not from my compiled character and ship stats. (Since I don't categorize M/M works, I couldn't do it any other way.)
Known bugs which affect the accuracy of this data:
- AO3 does not modify the "last updated" date of a work in a revealed exchange when the exchange reveals, so my database can miss exchange works entirely if they're last edited awhile before the exchange works are revealed. This probably leads to a serious undercount of works in revealed exchanges.
Share and enjoy, and let me know what I can do to make this more useful.
