Fandom mechanics; or, the lessons of GameFAQs applied to shipping and wank
I was going to write a long and spurious ship rating post for FFXII like I did last night for FMA, but then I got through the cutscene-o-rama with Punjabi Lando [1] and decided I had to digest that, so I went and read all the game mechanics FAQs on weapon stats, quickenings, bazaar goods etc. instead. I really wish I hadn't because now I have much more to worry about. XD;
However then I started thinking about applying the same modelling technique to fandom. That is, I was merely being precious last evening but then I figured, you could actually do something with this. XD
NOTA BENE: the following all relates to "passive" popularity, i.e. what proportion of people you could expect to like your pairing (or at least agree to entertain the idea), not what proportion of people are likely to write/draw for it. That introduces a whole slew of other factors (ex. size of fandom, how "difficult" the series/characters are to fic for, etc.), and besides I believe this last is random starting condition + power rule anyway. XD Similarly, the "wank" rating is a measure of how much tinder the warehouse contains, not whether anyone is going to light a match.
NOTA BENE 2: the following is pretty much bullshit, but IMO not more so than many economic models.
Canonicity (C): a lot of fandom meta confusion results from the fact that people talk about canonicity as if it were a yes/no binary, but it's not - at least it doesn't behave as such in a model of its fandom effect.
0 = "do they even meet?"
1 = "they know each other"
2 = "good friends/part of a tight crew"
3 = "a declared Special Connection"
4 = "other characters tease them about being a couple"
5 = "indisputably canonical girlfriend/boyfriend/married sort of deal"
IMO Western animanga fandom's gotten more rigorous about C as time goes on - 2 used to be plenty, now if people ship based on less than 3 they're faintly embarrassed about it and all "I'm on crack really!" I say to these people: remember that less than 3 spells ♥. (There are also older fandoms, like FF7, that operate to this day under a sort of grandfather clause, whereby if a ship's been around for long enough it doesn't matter that there's no real canonical support for it.)
Moe (M): the BL/het romance cliche/kink rating. A somewhat subjective mark given based not on how I feel, but how I believe the averaged-out animanga fandom would feel, based on lengthy experience therein. XD You could actually derive a formula for M by tabulating a loooong list of cliches and assigning them 1-3 points each based on how popular said cliches are (itself statistically compiled from that "bulletproof kinks" meme), then totting up a points total for each pairing and reading off a pre-determined scale for M, but I do not think manually going through this process would make the rating less bullshit than me assigning it by eye. XDD It would definitely be more impressive, though.
Note that it's perfectly possible for the cliches that go toward this rating to contradict each other, ex. 1x2 in GW fandom rates "genki retard uke" and "dewy-eyed longhaired angsty uke", simultaneously. But you could not assign it "master and servant" or "snarky partners in crime" or "best enemies". Also note that m/f or f/f automatically lowers M by 0.5 and 1 respectively - modelling reality, remember, not how it should be.
Personality/Meta (P): how well you could support the pairing in an argumentative essay based on characterization. 0 = "canon might say so but there's no chemistry and they're terrible for each other besides", 5 = "they are perfect for each other and here is the bullet point list why".
I finally decided that this is personality and meta, not strictly meta about personality - I just feel there should be a scale for the more intellectualized / manifesto end of ship justification, like TezuFuji (which most people would agree has a surprisingly low moe rating for a fandom juggernaut XD), or desire vs. self-realisation (XD), or the Xtreme Meta school of pairing generation whereby one comes up with a clever what-if first then sees how it could be done (ex. "If FFXII is basically Star Wars, wouldn't that mean [insert crack pairing]?"), as opposed to M which is just fandumb through and through. P is not weighted as heavily as M in the calculation though. XD
Controversialness (W): how likely it is other fans will hate your pairing, based somewhat subjectively on its intrinsic qualities (stuff like, is it underage/incestuous/non-con/on crack/destroys a widely adored ship just by breathing). Higher than 3 means a fair chance of wank. XD
Inherent controversialness discourages people from shipping: all else being equal, fewer people will ship something if the characters happen to be closely related, or their friends are vocal about disliking the pairing, etc.
The formula for passive popularity (PP):
PP = C + M + 0.5P - 0.3W
In real (fandom) life, moe weighs as much as canonicity and in fact may replace canonicity altogether. XD You can derive this yourself but the highest possible score is 12.5 and the lowest is -1.5 (maybe I should add 1.5 as a constant so the scale begins at 0?). Practically speaking SxS is 11.2, BanxGinji is 10.2, RoyEd is 6.5.
The formula for wank potential between shippers of pairings A and B, assuming A and B contradict (WPab):
WPab = 1 / $(PPa - PPb) + (Wa + Wb) + G, where:
$(PPa - PPb) = larger(|PPa - PPb|, 0.1)
G = 3 if gender(A) != gender(B), otherwise G = 0
...Okay I've just lost half of you. XD Basically the idea of this formula is, the more similar the popularity ratings of A and B, the more likely that corner of fandom will "split down the middle" and start rivalizing (inverse operation). Then you add the inherent controversialness of the respective pairings, plus a constant if good ol' slash vs. het is invoked. The rigmarole with the $-sign operation is just to avoid dividing by zero or getting a negative result.
Actually I think the curve is more complex than 1/x - there ought to be a way to account for the fact that 60/40 splits are more inherently controversial than 50/50 - but for a basic model along the lines of supply/demand sans elasticity it'll do.
The formula for total shipping wank potential of a fandom (WPt):
WPt = sum(WPab) * sum(PP) / (N * F)
Where N is number of pairings widely agreed to be representative (extreme fringe/crack pairings tend not to influence WPt), sum is over pairings included in N, and F is some cosmetic constant (probably divide by 1000).
I'm far too lazy to calculate WPt for anything, but a quick scribble shows WP(Ban/Ginji, Ban/Akabane) to be 4.18, WP(Roy/Ed, Ed/Winry) to be 8.53, and - as an example of why I need to improve the WPab formula re: 60/40 splits - WP(SxS, SxK) to be as low as 3.45. XD 4.37 - let's all pretend I can count. Note however that WPt for TB/X is high, because sum(PP) is ridiculously high for a relatively low number of total pairings N, i.e. emotions are more at stake than usual.
Suggestions etc. are welcome. XDD
EDIT, ABOUT AN HOUR LATER -- I think I'm going to change this so as to put the PPa + PPb term in WPab instead of (or perhaps in addition to) sum(PP) in WPt. That is if both pairings are very popular in the fandom it's much more likely there will be wank between them than if both are fringe ships, for example R/H vs. H/H is likely to cause wank, Ron/Luna vs. Ron/Tonks, not so much.
So the new formula is:
WPab = (1 / $(PPa - PPb) + (Wa + Wb) + G) * (PPa + PPb), where:
$(PPa - PPb) = larger(|PPa - PPb|, 0.1)
G = 3 if gender(A) != gender(B), otherwise G = 0
Going by this WP(Ban/Ginji, Ban/Akabane) = 77.562, WP(Roy/Ed, Ed/Winry) = 87.85, and WP(SxS, SxK) = 83.9. Out of curiosity I did WP(Ron/Luna, Ron/Tonks), and it gave 18.146. My instinct is that WPab is logarithmic but I wouldn't be able to prove that mathematically. XD
EDIT 2 -- figured out how to shift the curve in the above formula to account for the 60/40 effect. $(PPa - PPb) should be something like larger(||PPa - PPb| - 2|, 0.2)? Will crunch more numbers when I have the gumption, or better yet you can do it for me. XD
[1] All the mooglecraft tech support gets outsourced to Bhujerba these days OH GOD I'M SORRY I'LL JUST BE OVER HERE [2]
[2] And anyhow it's not as funny as "How come you always see a Hume guy with a Viera girl and never the other way around?" Thanks a million, fabulous_papaya!
However then I started thinking about applying the same modelling technique to fandom. That is, I was merely being precious last evening but then I figured, you could actually do something with this. XD
NOTA BENE: the following all relates to "passive" popularity, i.e. what proportion of people you could expect to like your pairing (or at least agree to entertain the idea), not what proportion of people are likely to write/draw for it. That introduces a whole slew of other factors (ex. size of fandom, how "difficult" the series/characters are to fic for, etc.), and besides I believe this last is random starting condition + power rule anyway. XD Similarly, the "wank" rating is a measure of how much tinder the warehouse contains, not whether anyone is going to light a match.
NOTA BENE 2: the following is pretty much bullshit, but IMO not more so than many economic models.
Canonicity (C): a lot of fandom meta confusion results from the fact that people talk about canonicity as if it were a yes/no binary, but it's not - at least it doesn't behave as such in a model of its fandom effect.
0 = "do they even meet?"
1 = "they know each other"
2 = "good friends/part of a tight crew"
3 = "a declared Special Connection"
4 = "other characters tease them about being a couple"
5 = "indisputably canonical girlfriend/boyfriend/married sort of deal"
IMO Western animanga fandom's gotten more rigorous about C as time goes on - 2 used to be plenty, now if people ship based on less than 3 they're faintly embarrassed about it and all "I'm on crack really!" I say to these people: remember that less than 3 spells ♥. (There are also older fandoms, like FF7, that operate to this day under a sort of grandfather clause, whereby if a ship's been around for long enough it doesn't matter that there's no real canonical support for it.)
Moe (M): the BL/het romance cliche/kink rating. A somewhat subjective mark given based not on how I feel, but how I believe the averaged-out animanga fandom would feel, based on lengthy experience therein. XD You could actually derive a formula for M by tabulating a loooong list of cliches and assigning them 1-3 points each based on how popular said cliches are (itself statistically compiled from that "bulletproof kinks" meme), then totting up a points total for each pairing and reading off a pre-determined scale for M, but I do not think manually going through this process would make the rating less bullshit than me assigning it by eye. XDD It would definitely be more impressive, though.
Note that it's perfectly possible for the cliches that go toward this rating to contradict each other, ex. 1x2 in GW fandom rates "genki retard uke" and "dewy-eyed longhaired angsty uke", simultaneously. But you could not assign it "master and servant" or "snarky partners in crime" or "best enemies". Also note that m/f or f/f automatically lowers M by 0.5 and 1 respectively - modelling reality, remember, not how it should be.
Personality/Meta (P): how well you could support the pairing in an argumentative essay based on characterization. 0 = "canon might say so but there's no chemistry and they're terrible for each other besides", 5 = "they are perfect for each other and here is the bullet point list why".
I finally decided that this is personality and meta, not strictly meta about personality - I just feel there should be a scale for the more intellectualized / manifesto end of ship justification, like TezuFuji (which most people would agree has a surprisingly low moe rating for a fandom juggernaut XD), or desire vs. self-realisation (XD), or the Xtreme Meta school of pairing generation whereby one comes up with a clever what-if first then sees how it could be done (ex. "If FFXII is basically Star Wars, wouldn't that mean [insert crack pairing]?"), as opposed to M which is just fandumb through and through. P is not weighted as heavily as M in the calculation though. XD
Controversialness (W): how likely it is other fans will hate your pairing, based somewhat subjectively on its intrinsic qualities (stuff like, is it underage/incestuous/non-con/on crack/destroys a widely adored ship just by breathing). Higher than 3 means a fair chance of wank. XD
Inherent controversialness discourages people from shipping: all else being equal, fewer people will ship something if the characters happen to be closely related, or their friends are vocal about disliking the pairing, etc.
The formula for passive popularity (PP):
PP = C + M + 0.5P - 0.3W
In real (fandom) life, moe weighs as much as canonicity and in fact may replace canonicity altogether. XD You can derive this yourself but the highest possible score is 12.5 and the lowest is -1.5 (maybe I should add 1.5 as a constant so the scale begins at 0?). Practically speaking SxS is 11.2, BanxGinji is 10.2, RoyEd is 6.5.
The formula for wank potential between shippers of pairings A and B, assuming A and B contradict (WPab):
WPab = 1 / $(PPa - PPb) + (Wa + Wb) + G, where:
$(PPa - PPb) = larger(|PPa - PPb|, 0.1)
G = 3 if gender(A) != gender(B), otherwise G = 0
...Okay I've just lost half of you. XD Basically the idea of this formula is, the more similar the popularity ratings of A and B, the more likely that corner of fandom will "split down the middle" and start rivalizing (inverse operation). Then you add the inherent controversialness of the respective pairings, plus a constant if good ol' slash vs. het is invoked. The rigmarole with the $-sign operation is just to avoid dividing by zero or getting a negative result.
Actually I think the curve is more complex than 1/x - there ought to be a way to account for the fact that 60/40 splits are more inherently controversial than 50/50 - but for a basic model along the lines of supply/demand sans elasticity it'll do.
The formula for total shipping wank potential of a fandom (WPt):
WPt = sum(WPab) * sum(PP) / (N * F)
Where N is number of pairings widely agreed to be representative (extreme fringe/crack pairings tend not to influence WPt), sum is over pairings included in N, and F is some cosmetic constant (probably divide by 1000).
I'm far too lazy to calculate WPt for anything, but a quick scribble shows WP(Ban/Ginji, Ban/Akabane) to be 4.18, WP(Roy/Ed, Ed/Winry) to be 8.53, and - as an example of why I need to improve the WPab formula re: 60/40 splits - WP(SxS, SxK) to be as low as 3.45. XD 4.37 - let's all pretend I can count. Note however that WPt for TB/X is high, because sum(PP) is ridiculously high for a relatively low number of total pairings N, i.e. emotions are more at stake than usual.
Suggestions etc. are welcome. XDD
EDIT, ABOUT AN HOUR LATER -- I think I'm going to change this so as to put the PPa + PPb term in WPab instead of (or perhaps in addition to) sum(PP) in WPt. That is if both pairings are very popular in the fandom it's much more likely there will be wank between them than if both are fringe ships, for example R/H vs. H/H is likely to cause wank, Ron/Luna vs. Ron/Tonks, not so much.
So the new formula is:
WPab = (1 / $(PPa - PPb) + (Wa + Wb) + G) * (PPa + PPb), where:
$(PPa - PPb) = larger(|PPa - PPb|, 0.1)
G = 3 if gender(A) != gender(B), otherwise G = 0
Going by this WP(Ban/Ginji, Ban/Akabane) = 77.562, WP(Roy/Ed, Ed/Winry) = 87.85, and WP(SxS, SxK) = 83.9. Out of curiosity I did WP(Ron/Luna, Ron/Tonks), and it gave 18.146. My instinct is that WPab is logarithmic but I wouldn't be able to prove that mathematically. XD
EDIT 2 -- figured out how to shift the curve in the above formula to account for the 60/40 effect. $(PPa - PPb) should be something like larger(||PPa - PPb| - 2|, 0.2)? Will crunch more numbers when I have the gumption, or better yet you can do it for me. XD
[1] All the mooglecraft tech support gets outsourced to Bhujerba these days OH GOD I'M SORRY I'LL JUST BE OVER HERE [2]
[2] And anyhow it's not as funny as "How come you always see a Hume guy with a Viera girl and never the other way around?" Thanks a million, fabulous_papaya!