• Member Since 11th Apr, 2012
  • offline last seen 10 hours ago

Bad Horse

Beneath the microscope, you contain galaxies.

More Blog Posts691


Some statistics on fimfiction.net: Ranking authors · 5:37pm May 14th, 2012

This weekend, I wrote a spider to suck down and parse all the pages on fimfiction describing the stories. Mostly I looked at story title, author, number of views, thumbs up and down, genre tags, and character tags.

There were 10,590 stories with 39,846 chapters by 5,512 authors, having 9,618,891 first-chapter views, 368313 thumbs-up, and 35130 thumbs-down.

Like money in America, most of the chapter views go to just a few authors. The most-popular 274 authors (5%) got half of all views. The most-popular 71 authors (1.3%) got 1/4 of all views. This graph shows the number of first-chapter views each author got, arranged from fewest at the left to most at the right. It shows why I think the "featured stories" box should be eliminated - it steals from the poor to give to the rich.

Top authors ranked by popularity are:

AUTHOR rank stories views thumbs
ROBCakeran53 1 2 191641 0.011
AbsoluteAnonymous 2 19 77787 0.065
Blueshift 3 16 75190 0.014
FanNotANerd 4 6 74157 0.011
theworstwriter 5 9 69362 0.009
JasonTheHuman 6 3 60479 0.032
Ebon Mane 7 12 60026 0.011
Cold in Gardez 8 9 58856 0.028
midnightshadow 9 25 53794 0.001
PotatoJoe 10 14 43362 0.019
DawnFade 11 10 41037 0.033
butterscotchsundae 12 10 39341 0.035
Thyrai 13 4 38069 0.020
Andrew Joshua Talon 14 7 37627 0.047
JohnPerry 15 11 37105 0.011
Vargras 16 7 36733 0.037
Paleo Prints 17 5 36385 0.022
Dubs Rewatcher 18 11 36280 0.020
StreakTheFox 19 7 34981 0.005
Carmine 20 17 33478 0.018
Tailslover13 21 6 33273 0.010
Samsara 22 5 32765 0.023
Chatoyance 23 14 32075 -0.011
Conner Cogwork 24 6 32017 0.056
Squeak-anon 25 12 31856 0.032

But how much is view count a measure of quality? Even if it's good, it isn't very useful, since you already know who the popular authors are. Are there authors who may or may not get very many views, but get great ratings from the views they get? Yes, there are.

The 'thumbs' column is a net-thumbs-per-view rating. I rated authors by the score (thumbs-up - 9.0624 * (thumbs-down + 1)) / views, which makes the average score zero. It turns out that nice, uncontroversial, canon-like stories like "Ah Ain't Got No Ack-cent" do well by this score, but disturbing or risky stories like device heretic's "So Be It", which get a lot of thumbs-downs, don't do very well. It favors less-popular authors, because it's easier for them to get a good score by chance due to having fewer ratings. This made it useful for finding wonderful and obscure stories, like Hadles' Near Death Experience, Moonspire Run by titanrising, and Directive: Grow by Dragon Dreaming. But it also let some mediocre stories score very highly, because they had some thumbs-ups and no thumbs-downs yet. So I made a blend of views and thumbs-per-view, which is

thumbs-per-view-score * ln(views)

(If I were less lazy, I might have used the z-value of net-thumbs-per-view instead, which tells you how many standard deviations away from the average score they are after adjusting for sample size.) Top authors by this score are

Hadles 776 1.344
Pen Stroke 8789 1.280
Ciroton 1174 1.274
Visiden Visidane 1298 1.203
Capn_Chryssalid 3474 1.190
PhantomFox 13876 1.140
Pegasus Rescue Brigade 6203 1.129
InsertAuthorHere 10379 1.106
Thanqol 1037 1.096
Chromosome 4328 1.091
Shadowsca 302 1.020
zaptiftun 5807 0.987
Denim_Blue 2744 0.974
determamfidd 7913 0.932
Mechcolt 112 0.922
Peroth E. 1018 0.917
Razorbeam 4749 0.885
TeaPartyCannon 6810 0.873
shortskirtsandexplosions 18803 0.862
Cupcakes 6600 0.848
treg388 596 0.835
Nyerguds 1934 0.801
TAW 7026 0.798
Municipal Engines 5067 0.789
Explodium 2241 0.787
DeadParrot222 407 0.781
koolerkid 930 0.771
device heretic 19173 0.769
boardgamebrony 7270 0.766
Aldea Donder 2840 0.760
Krizak 5013 0.748
Rust 8952 0.742
AbsoluteAnonymous 77787 0.729
Flower Seven 473 0.728
CardsLafter 6322 0.728
CoffeeGrunt 14464 0.724
Altoid 1411 0.719
Gravekeeper 13002 0.718
Bad Horse 1019 -0.043 :rainbowhuh:


Sadly, you can only do so much working with ratings from all readers, because readers have differing tastes. Also, when a story is really good, users are more likely to Favorite it; and when they favorite it, they often don't bother to give it a thumbs-up. Many stories have been favorited twice as many times as they've been given a thumbs-up. So, there are a few bad authors on this list. There are some very good authors, like Skywriter, who don't rank well on it (#333).

Next time, I'll rate different stories on their grammaticality, and find out how important grammar is to readers.

Report Bad Horse · 2,756 views ·
Join our Patreon to remove these adverts!
Comments ( 9 )

Have you started to do breakdowns about user habits? Like, people who like this story also like this one, regardless of story's net popularity?

Wish I could, but I don't have the data on which users gave thumbs-up or thumbs-down to stories.

Can you do it by who's favorited? Or do the thumbs-up style data with favorites?

This is impressive, I just have to say.

I could scrape the site for user favorites and author tracks, if I had a list of user names. I don't have a list of user names, though I have a list of author names. It would be a bit more work. It would be a lot easier if I could get the data from knighty, but I haven't asked him - I don't know if he even checks his email, as he's never replied to me in the past.

You get the list of author names from the list of stories, right? If you have all stories, you can back out to the list of all users that have favorited a story by scouring the favorites that have been left on stories, right? And for analysis-by-favorites, you don't care about someone who's never left a favorite, anyway.

Pretty nice ending to an interesting analysis. :heart:

I was trying to find that post about statistics on posting times, and I found this.

I was the eighteenth most popular author on the site at one point? What the fuck? :rainbowhuh:

I miss AbsoluteAnonymous and Device Heretic. Considering their um... particular personalities, I might be the only one.

Login or register to comment
Join our Patreon to remove these adverts!