Writing to not be disliked · 2:32am Feb 10th, 2013
One of the many amazing things about fimfiction is that, no matter how awful a story is--even if it is the most-vile story ever written--many people will love it. So number of thumbs-up is a poor indication of quality. Being featured is a poor indication of quality. Lousy as it is, the next-best thing we have to reviews and advice is thumbs-up / thumbs-down. That's how the "top-rated" page (under BROWSE) is sorted.
To get on the the first page of "top rated" stories, you need almost 100 thumbs-up per thumbs-down. It's worthwhile to get on that first page! I have two stories there, Sisters and Mortality Report, and they keep bringing in new viewers every day.
But if you look at what's on that page--they're great stories, but for the most part, they're not controversial or edgy. You don't usually get 80 thumbs-up per thumbs-down by being edgy. Pissing one reader off because you said Vegemite was disgusting cancels out 1000 readers who liked the story (because only 100 will bother to give it a thumbs-up). Having 2 or 3 readers who hate you enough to follow you and down-thumb everything you write would probably keep you off that top-rated first page forever.
This is the future of fiction. In the past, we read fiction that got some great reviews, or was on a best-seller list. Today, I buy my books from Amazon and the first thing I do before buying is to look at the negative reviews. In the Amazonian age, having just one reader who hates you can seriously damage your writing career. Having one arbitrary element that triggers a knee-jerk reaction in 0.1% of the population would probably double the number of 1-star reviews on your book's Amazon page.
I've found myself thinking too often about the thumbs-down. On Pony Tales, I got suddenly hit with a lot of thumbs-down when I posted chapter 4, "Shipping Costs", until I figured out that it was people who were butt-hurt because the story makes fun of My Little Pony generation 3. I reworded it to be kinder to G3, and the down-voting stopped.
Think about that. I went back and changed a story because 3 people out of a thousand were enraged enough about disrespect to Gen3 to hit the thumbs-down. Because of how our ratings work, avoiding pissing off a tiny fraction of readers for stupid reasons is more important than improving the story for the other 99.9%.
By now I know some of the thumbs-down triggers. Non-canon, alternate-universe treatments of characters will always get a lot of thumbs-downs, even if you tag the story "alternate universe". Foul language. Evil Celestia. And (sniff) red-and-black alicorns.
I knew that the latest Pony Tales chapter, "The ones who walk away from Equestria", would attract down-votes. I even put a warning at the top to keep butt-hurt-prone people from reading it. (I don't especially recommend you read it. My watchers have suffered enough at my hooves already.) It's been up 1 hour, and in that time Pony Tales has gotten one-fourth of all the down-votes it's gotten in the past 10 months.
Fortunately I can just say "screw it" and push onward in this case, because it doesn't matter much how many down-votes that story gets. It's not anywhere near getting on the first page of top-rated, or even page ten. But if I were writing for a living, it would matter, and I would avoid writing stories that made people uncomfortable. Unless we find a better way to rate and recommend stories, writers will more and more stick to writing nice stories that don't inspire hatred or other unruly emotions, because modern technology has made the voice of one disgruntled moron louder than the quiet hum of a thousand contented readers.
I'm not complaining about having the thumbs-down button. It's important! Nor am I really concerned with fimfiction. I'm thinking more about Amazon and other e-pub outlets. Fimfiction is the future of publishing. Readers want to be part of a community, and interact with authors and with each other. It's going to be great. But it will change things, too, in unpredictable ways.
ADDED: I think the best solution is to encourage people to down-vote more frequently. Then down-votes wouldn't be dominated by a few cranks.
The only thing I could think of would require all votes to have some text reasoning behind them, followed by a review by the site to make sure it's a valid complaint and not whining.
I would propose a sub-argument there: This is true of general fiction, but not of genre fiction.
In the world of ponies there's really no such thing as the sci-fi and fantasy genre, and we have no tags for mystery, and I can't speak for dark/gore/horror. However, I happen to know a lot about the romance genre. Ship fics of popular ships get automatic thumbs down, before it's even possible for people to have read them, (usually about ten, depending how long they're on the front page, and you can double that if it makes the feature box or contains clop.) I guess it's our version of the "genre ghetto" that real authors talk about. But I get a handful of faves daily on my romance stories, because people actively look for romance stories in their favorite shipping. And even though I've published nothing new in over a month I still get followers every day or so, because I'm a shipping author.
In real publishing you have a similar situation. There are networks for sci-fi and fantasy fans- blogs that review just these books, magazines or websites that publish short stories, and fans who devour those kinds of books specifically, buying and reading armfulls at a time. The same is true of mystery, horror, and romance. Really, if you want to be a successful author these days, genre fiction is the way to go. You probably won't win that Noble prize for Literature, but you'll get the readers a lot more easily than general fiction, and they'll follow you more closely.
I'm pretty sure playing it safe isn't a winning strategy either. I don't write 'edgy' to avoid controversy, the stories I want to tell just don't happen to be edgy. I'm pretty sure this has gotten me down thumbs for being bland. Or at least for being perceived as bland. Many readers search for stories based on warnings; having no warnings and being rated for 'everyone' will get you passed by.
A very interesting dilemma, indeed.
I've downloaded most of the stories I've liked over the years and keep them organized in Calibre, which has also allowed me to download and sort the metadata associated with stories. That includes thumbs up, thumbs down, number of views, number of words and chapters, etc. I've done some very simple statistical work with the likes/dislikes, using error bars to account for low sample numbers, but as one might expect that doesn't do much to change the sorting of the more popular stories since they've got very large samples. Perhaps I should try adjusting for views in some way.
But unfortunately, I think in the long run there's not going to be a good universal answer. There's tens of thousands of stories on this site being read by tens of thousands. The "top rated" ones among such a vast population are inherently going to be the ones that nobody hates. I think maybe the best thing to do is to try to facilitate the grouping of stories based on other criteria, rather than just having one single "top rated" page. FiMFiction made a little progress with the recent revamp that made the faves less prominent, and there's the groups mechanism that could perhaps be leveraged in this regard too.
I think you've hit a nail on the head here, and I've been wondering about this myself. My last two posted fics have done very well in terms of all-time ratings and naturally wondered why [1]. I actually had an operability to ask Knightly how the top-rated metric works, and it turns out to be based on the Wilson Interval somehow. Read about it here. So I guess all I did was write two fics that sufficient people liked and very few people disliked. (I didn't set out to do this, it's just how it ended up).
I wish I could offer a better solution, but trying to automatically sort stories (or nearly anything) mechanically in terms of goodness is really, really hard.
Hummm... One of the best indications of a good story is probably one that stays with the reader so they remember and recommend it. To that end, I wonder what rating stories based on how many times they get posted on peoples' user pages would show?
[1] I mean, it was flattering to see my story up there near the top, but on the other hand, I was above lots of stories I'm pretty sure I didn't deserve to be above.
812765
Would it be that difficult to implement a shipping censor similar to the no mature content option? It wouldn't do anything against the people who make it their personal mission to downvote all shipping as a matter of principle, but it would clear up the people who are just "get these darn lesbian horse stories outta my feature box!"
As for the rest of the problem, you'll never be able to entirely fix the popularity =/= quality issue. People in groups will often like things that, from a subjective or technical perspective, just aren't that good. People have bad taste sometimes, and people in groups have worse taste that tends to speed like the flu.
There are a number of ways to get around this. You could set up a more in depth reviewing process to detail a story strengths and faults. This would have the added bonus of deterring impulse down ores by making the process take longer, and you could somehow aggravate the reviews to point to better stories, either by number of reviews or standing of the reviewer. There a lot of possible metrics you could work with depending on what data you're set up to collect. It'd probablly be as much of a pain to implant as the "people who liked this also like story X" ideas I've heard batted around, but its something to think about.
My solution? Keep on following that list till the dislikes start outnumbering the likes. Partially because if I didn't, I would run out of stuff to read (especially ship fics; I didn't realize the extent back when you could filter out stories marked as read, but I have most of the first 90 pages of fics with the romance tag marked like that, indicating that I have either read it or confirmed that I don't want to read it). Mind you, I read a massive amount more than most people (Back in high school, used to go through a Heinlein or Clarke novel every day or two), so that's probably only related to my situation.
When it comes to review based sorting, I tend to ignore reviews if I don't already agree with the reviewer in some way (SF, they have to be able to quote Asimov, Heinlein, or Clarke. Fantasy has to be Xanth or Incarnations of Immortality), or if they have a good track record. I agree its an issue, but the solution isn't anything more than "get them to read more". If someone only reads the equivalent of one novel a year (can't even understand such a low rate), then of course they're going to pay more attention to the negative reviews. They evidently don't have much time, so why would they invest however long it takes them to read something which they have reason to think they will hate? If someone is running out of books to read, of course they're going to care less about the negative reviews, if only because they've most likely read everything which never gets negative reviews.
Ouch. So a very small number of *-holes can reduce your writing to the dustbin of ignored stuff. I didn't realize the impact of (-) votes, just been looking at the (+) thumbs. Out of all my stuff, I only have one (In Celestia We Trust) with a -/+ ratio of 1:100 or better. I first thought I got a number of (-) votes for my Twist dialogue in Monster, but the vast majority (8/12) of them came *before* her chapter. The (8) probably don't like the idea of Trixie as an Element Bearer. Even with 600+ likes, those 12- votes have kept Monster off the charts except for that first wonderful week where my ego blew up to unimaginable proportions.
I don't think it's possible to write an enjoyable fic that doesn't irritate somebody. I still think my two "romance" fics got a number of (-) from the "Where's the clop" crowd. I'll agree that the real tack under the saddle is the downvoters don't even have the courtesy to sign the little bags of flaming poo they leave on the virtual porch, they just fling and fly, leaving us to say "What did we do wrong?"
The downvote button is not useful. I say it is not merely pointless, but ultimately toxic. It serves only to force writers to write to the lowest common denominator, and it does not reflect the real world.
In the real world there are no downvotes - people only upvote with their wallets. They buy, or walk away.
Having only an upvote button, and NO downvote button is like having a shop where the reader can buy, or walk away. They like it, or they do not, plain, simple, and - according to your figures - one hundred times less prone to abuse.
The downvote button does nothing but teach that blandness is the key to success. It should be abolished.
812813
(reposted because I forgot to tag alexnine)
I don't know how hard it would be, you could suggest it to Knighty.
I'm also not sure how much good it might do. In addition to the rabid anti-shippers, and the less-rabid but still downvoting anti-shippers, there are a number of people who actually do ship, they just don't like this ship. I suspect quite a few of them, because uncommon ships (like my Daring Do x Zecora fic) get fewer auto-downvotes than popular ships (like my AppleDash stuff.) Also, a small handful of people on my survey, people who already claim to like shipping, have admitted to doing this to ships they dislike. So I think there will always be some problem with auto-downvotes in shipping.
There will always be a problem with auto-downvotes, unless of course...
We get rid of downvotes!
We get a much more informative readout if we simply have an upvote-no upvote system. People can look at how many people read a story, and how many of those liked it enough to click the button. It turns the rating from an overly complex trinary system, into a simple binary system. Rather than confusing middle grounds, readers would know that X% of viewers found this story worth the time it took to read it.
It'd even make calculating which stories to feature require less computation! Because the algorithms would be working with one less variable.
Not just literature really. The internet has skewed opinions like that for damn near everything.
Still, on the other hand, there is no question that the internet also allows people the ability to reach an audience that far outstrips what was possible before. Just look at this site. It's got a rather massive following, consisting of both an enormous amount of authors, and even more readers. And this is just one fandom.
Still, it is going to be impossible to please everyone, and there are always going to be assholes. Hell, that was a truism long before the internet ever came about. I just don't know how one would fix it. I do wonder though, if this is part of why fewer and fewer books have a synopsis on the back. Instead they all seem to have "Praise for the author" all over them. No more telling a potential buyer, "here is my story, it's interesting no? Buy me." It's all, "Listen to other people tell you how awesome my story is, and buy me"
I really hate that, but given what you just said about one negative opinion, maybe authors and publishers feel they have no choice.
As an aside, I almost never hit the thumbs down, even if I don't like a story, because it's usually because of a personal problem with the story, not an issue with the quality of the story itself. If that makes sense.
812857 812927 You can look at a story now and see upvotes / views. I don't want to get rid of information. I'm pretty sure the featured box already doesn't consider downvotes much, because I've seen heavily-downvoted stories in the box. The top-rated stories page does, but if you rated things by upvotes without considering down-votes, the top-rated stories would be largely about Halo and Transformers.
I'm more concerned about bad reviews on Amazon and e-publishing sites than on fim-fiction, and on how this will affect whatever replaces the New York Times Bestseller List. Will bookstores stock books based on down-votes as well as on sales? Perhaps they won't--sales are what they care about; to them, down-votes are irrelevant. But bookstores are increasingly irrelevant. They are going to provide a fraction of the sales of physical books, which will be a fraction of the sales of books.
812872
This would be the main draw of a more in depth review process in my eyes; you could weed out the ligitimate negative criticism from the "AppleDash sucks Twidash forever" OTP downvoting. It'd be the same for a trusted reviewer system.
I think the process for a shipping censor would be fairly straightforward. Right now the mature censor hides all content sight the [mature] flag attached to it. All you'd need is a second check box that does the same thing for the [romance] genera tag. If you wanted to get really fancy with it, you could probablly build the system to exclude any of the listed tags. Don't like [dark]? Now you don't have to worry about it! Don't like [gore] but like [sex]? Done! Only interested in [adventure] and [slice of life]? Everything else is now no longer your concern! My brain is probablly dramatically oversimplifying how easy this would be to code, but at least it seems possible, since its just an extension of something that's already in place. A working detailed review system would be a whole different nightmare.
>caring about downvotes and fic rankings
Bad Horse, you crazy
In the end you just have to write for yourself, over thinking or writing about things you don't believe is just selling yourself short. Nobody wants to be hated especially over something trivial but as long as you can be honest with yourself your work will be appreciated for what it is.
813333
I'm more concerned about how user ratings will affect all of literature. I think it's not a big problem on Amazon presently, because users have to write a comment with their rating, and you can usually figure out which ratings are valid (at least for the bad ones).
Besides, recognition is nice, y’know?
812798
Thing is, the goodness of a story is heavily subjective. We can talk about good writing, sure enough, but what people want when they look at a rating is "Will I like this story?"[1]. And depending on any given user of the site the optimal method for figuring this out is different. If it's an anonymous reader-lurker then upvotes have to matter, because there's little else. But for me, specifically, I'd much prefer if FiMFiction rated stories for me by paying close attention to what my friends have read and upvoted. Five hundred upvotes on your story? Brilliant. I may just read it. Upvotes from the likes of Skywriter, or our illustrious host here, or, indeed, you, PoweredByTea? I have to read it now.
[1] I'm picky about my stories (it takes surprisingly few off turns of phrase or similar blunders for me to leave halfway through the first chapter) but even I care most about what I'll enjoy.
812957
No, no, it does make sense. I rarely make use of the downvote button, myself. Only for fics that offend me, basically. Things I'd describe as not merely bad or for someone unlike me or something I'm simply not into, but actually foul. The...thing our illustrious host linked to being a prime example. A synopsis made me ill.
812970
I don't know if I'm representative (I suspect I'm not) but upvotes/downvotes are of limited utility for me when shopping for books. I go by writer first, if it is a writer I like, I tend to buy impulsively. If it isn't, my next recourse is word of mouth, and after that actual reviews from sites I halfway trust. How many stars Amazon says it's worth? That impacts me surprisingly little. I think it comes from my long years playing video games, where review scores are actually worse than useless.
813333
He's a Hoardsmith. Of course he's crazy. But in this instance he's not as crazy as you might think. Aside from the complex mess of ratings, upvotes, downvotes, featured boxes and such being an interesting system to study and think about if, ah, well, if you like studying systems, there's another dimension to consider. Publishing (the mainstream, not about ponies kind) is changing and how fanfiction works may be a useful model to predict how some of those changes will look like. Ratings being a prime example.
813425
*Reads synopsis of A Fun Day*
memeorama.com/wp-content/uploads/2012/06/nop-nope-octopus.gif
813425
Perhaps. I very much doubt mainstream publishing will ever relax to the level we're at, though, so the exercise feels more like particularly imaginative, but largely moot, futurist musings to me.
Good essay.
But A Fun Day the most vile story ever written? LOL. NO. Not by far. Not even by far within FiM fanfics.
813495
I swear, it's like telling someone not to look down. First damn thing they do.
For the benefit of later readers: Don't look up...that story. Don't. Seriously. Nothing good awaits you at the end of that particular rainbow.
*sigh*
On the plus side, OCTOPUS! A Mimic Octopus, I think. Those are awesome.
...what? I like cephalopods.
813503
As for relaxing, no I suspect it won't. But both future publishing and current FiMFiction-uh-ing share a problem with ranking and grading books. Hence is useful to think about the former in terms of the latter. Well, possibly.
That being said, you could be right. But really, particularly imaginative, but largely moot, futurist musings are their own reward.
I think the future of media recommendations — literature, music, games, etc. — is based on social networking. Please don't kill me.
I follow Bad Horse and I read stories that he likes because I think I'll enjoy them as well. Like 813425, I rely on recommendations from friends and favored reviewers. Say I had a way to find common favorites among like-minded individuals. That's how I would search for good reads.
812857 and 812927 bring up the idea of removing downvotes, which would better reflect actual markets. This is an interesting idea that I have seen implemented on EqD and certain subreddits. The biggest issue, as 812970 points out, is how controversial stories and comments are ranked. If we used an upvote-only system on this site, our top-rated stories would be about Son Goku the alicorn using his giant sword to defeat Generic Vampire Lord #12 with his buddies Sergeant Naruto and Pikachu Sonicsson.
The fact of the matter is that different groups have different tastes. The best way to make everyone happy is to let them choose whom they want to associate themselves with. Voluntary social networking + locating commonalities = success.
When I buy something on Amazon, the first thing I do is check the downvotes — to see if they are articulate and rational, or whether they're a flat "this sucked." I'm so used to the Disgruntled One Percent Effect that I only start worrying if the one-stars are a significant percentage or if they are pointing out a problem I'm not willing to overlook.
The DOPE[1] is magnified on FIMFic, because there's no mechanism for assigning relevance to the opinions of the crowd; this site has a relatively naive dislike-expression system (often there's no way to find what it is that people dislike without reading the story anyway), and a relatively naive top-story-sorting algorithm (as you describe). This is not to say it's not a problem! However, I'd want to see more evidence before generalizing from FIMFic to the writing universe, where better filtration tools exist.
--
[1] Swear to Celestia I didn't plan that acronym.
If we keep up and down votes as they are, we get bland quality.
If we only keep upvotes, we get bland garbage.
I don't think the problem here actually lies with downvotes.
813728 On the plus side, OCTOPUS!
Octopus? Wait, is Pus one of the hospital patients? Or maybe you mean Octaopal?
Oh, I remember. It's a word.
All in favor of moving to a system that only has downvotes?
814254 Do we agree by down-voting your comment?
813732
I totally agree with this. In fact, the small press publisher producing my original fiction novel (Oloris) is a branch of a social networking site (the Middle-Earth Network), and I think that's a pretty brilliant way of doing business- the small press has built in, targeted marketing for its products, and becomes a money making side project for its parent company, which can remain focused on fanworks and content.
Additionally, I suspect I'll get a handful of sales from my activities here when the book is published. While my book has nothing to do with ponies, people who read my stories and my blog know that they like me as a writer, so if it sounds interesting to them they'll be more likely to take a chance on it.
This actually brings up another thought I had for Bad Horse- how many of those Top Rated Stories are on EqD? I suspect a healthy percentage. In those cases, they didn't become top rated solely based on perceived quality, but on a combination of popularity that comes partially from the approval of a source the audience trusts, combined with a lack of offensive qualities that turn off potential readers. Add to that the number published by authors who already had a large following, even if they weren't on EqD, and I think most of those benefited from a number of factors besides quality and blandness.
It's always been the case that new, or unconventional literature has succeeded or failed based on word of mouth. But now more than ever, when people can group themselves based on interests and seek out voices from all over the world to find opinions they trust, and in a world where "browsing" becomes difficult, word of mouth is going to be the deciding factor in getting creative projects off the ground. There's a lot of "career building" that goes into being successful, both here on FiMfiction and in real publishing, but networking is the key to that.
Proposed: A new ranking system option that can be summarized as "Favorites of Followed" much as the "Degrees of Kevin Bacon" game.
In this ranking system (Let me pick on GoH, BH, and Jmac for an example) let us say I'm following these three, and Jmac sees a new fic named "Bob" and gives it a favorite and a thumbs-up. I click on my (proposed)FoF search on FimFiction, and it shows the new fic with a ranking of .333. Later GoH gives it a fave just to keep reading it, and the ranking I get from it would jump to .5 or so. Then BH looks at it, and gives it a thumbs-down, dropping its ranking in my search back down to ______ (not sure what weight to give Faves from Followeds vs Downvotes from Followeds). Later all three of them Fave and Upvote a fic called "Fred", which will show up in my search as a 1.0, indicating an unread fic that I really should put on the top of my reading pile.
A) Does this seem implementable? and
B) Does this seem like it would be useful? The Transformers fans would tend to Follow other Transformers fans, and their picks would float to the top of their searches, while leaving the HiE fans to their cliques, and so on, and so forth....
This is a little similar to my favorite way to look for good fics. I go into my "followed" group, and search through their favorites when I"m feeling bored.
814459
I don't know what the internals of FiMFic are, but it seems relatively simple to implement.
There's one problem with such systems, of course. They lead to small cliques doing their own thing, but a new reader being in a pickle because there's no good way to get good stories. Your recommendation system is only as good as your social graph - and if you've just joined you are cast adrift on a sea of Transformes crossovers.
814254
That...sounds hideous, actually.
814632 Well, new readers still have the regular "upvote/favorite" metric to use to determine good stories, and as they determine what is "good" for them, the computer algorithm becomes self-correcting. (although it would suffer from selection bias as the reader becomes isolated from "good" fics in other reading preference areas)
813732 814459 814632 I'm working on a similar system. The first approach I used just uses nearest-neighbor clustering on favorites. The second, which is best algorithm known for recommendations based only on user-rating pairs, uses singular vector decomposition to construct feature vectors for stories and users as linear combinations of favorites vectors. Haven't done anything with watchers yet.
Unfortunately the site upgrade broke all my code, and the Java Server Pages protocol is intimidating, and I haven't had time to work on it for months.
814402 >This actually brings up another thought I had for Bad Horse- how many of those Top Rated Stories are on EqD?
More than half. I checked the first few dozen once.
>It's always been the case that new, or unconventional literature has succeeded or failed based on word of mouth.
I find word of mouth doesn't work at all. Here, you get read by being in the feature box, or on EqD. I have 50,000 story views, and I think at least 30,000 of them came directly from EqD or from the featured box. Most of the rest were from people who read a story of mine thru EqD or the box then reading another story of mine. Burning Man Brony got great reviews on One Man's Pony Ramblings and by Seattle's Angels, which are the big two of pony reviews AFAIK, yet still has only 755 views.
On average, if 100 people read a story, 10 will like it, 10-20 will favorite it, and 1 will watch the author. (The longer I've been around, the fewer watchers per view I get, as more and more of my readers are people already watching me. Corpse Bride got almost a thousand views, and I got one new watcher from that story.) For word-of-mouth to create a sustained chain-reaction, each new reader has to convince 100 other people who've never read my stories to try one. That's not even close to happening.
814632
In fairness to me, it was kind of a joke.
814764
This is based on a fallacy. Those numbers are, as you say, on average. Given all of the places people might hear about your fics- random browsing, the front page, the feature box, EqD, and word of mouth- that is most likely true, or at least I'll happily grant that to you since I never looked into it myself.
The fallacy comes in assuming that the 100 people who, for example, see your fic in a blog post I make are going to act the same as a random sample. They're already separated out by a factor- they like my stories and agree to some extent with my taste. So if I liked your story, and 100 of my followers check it out, the like rate will probably be above average because the sample is skewed towards "liking things similar to what I like."
That sample can be skewed ever farther. If someone who shares my sense of humor says they're looking for a parody of alicorn OCs, and I direct them to your story, the chances of them clicking like are far greater than 1/100. Word of mouth is more likely to be targeted marketing, people don't just go around telling 100 random people to check out your story, they tell 20 people that they have some connection with, and that connection makes it more likely that they'll know something about that person's tastes.
We could test this some time, if you like. One of us puts up a story, but doesn't post it. The other tells X number of people they know, who they think would enjoy this particular story. I'll bet you it gets more than one like.
812798 Your story wasn't just "near the top", it was AT the top for a while. It was number 1! Until some second jackass gave it a thumbs-down. I didn't mean to slight your story. I like it a lot!
I've actually noticed this impacting me a lot more than I expected it ever would. Honestly, I think that may be partially responsible for my recent writer's block, which I'm finally just getting past. Not that I want to change the content of what I write so much as that I'm concerned that what I am writing won't meet up with people's expectations. I need to learn to get past that, or I won't ever take any risks. And, if I never take risks, I might as well wrap it up right now and take up a new hobby. Paint-by-numbers, perhaps.
There is one fundamental problem, I think, with trying to implement any kind of change. Part of why the things that are top rated, and featured, and so forth, is that they are what the majority of people like. To paraphrase Vargras, we are a fandom consisting primarily of 20 something year old males, and this tends to be reflected in what is popular to read. Which often tends toward one shot comedies, and naughty bits. No amount of adjusting the ratings system will change that, unfortunately. (Not that there is anything WRONG with comedies or naughtiness ((lol wut?)), by any means, it's just a small portion of a whole, and is over-represented.)
819419 I think that's a separate issue. Even the opposite issue. I'm talking about how, in a system where people are reluctant to leave down-votes, the few down-votes that there are are disproportionately important, and give too much emphasis to the opinions of a few cranks.
The solution, I just realized, is to encourage more down-voting! The problem only occurs because people are so reluctant to down-vote, that one or two down-votes is important.
819473
Which would be difficult to do, since by now, everyone thinks of a down-vote as a BIG negative, thus are reluctant to down-vote.
819159
I know, and I have a screenshot saved. I was just wanted to avoid boosting about what might have been a statistical anomaly. And, um, I probably do share as many traits with Quizzical Greystone as that comment made it sound.
I don't feel insulted, because both my fics were exactly as I wanted to write them. I didn't change anything because I was afraid of dislikes, it just so happened that the two stories I wanted to write were extremely uncontroversial. I'd like to think they were still good though. I'm actually bracing myself for dislikes on the next piece I've just finished.
Second thumbs-down? It was also briefly at the top with 0 thumbs-down and 207 thumbs-up (this lasted about 24 hours). You mean it was at the top twice? The top place may be a bit of a curse, it would seem.
823708 Oops, no, that would've been the first thumbs-down that sent it down the list, then. I wondered about that.
Upvotes/ (views)/ downvotes? With added weight to repeat viewers (IE, someone who came back to read the story again)?
Of course, that only works with something like here, where you're tracking both views and repeat views. It wouldn't work on, say, amazon.
I've been thinking (since 814254 mentioned it) about what a downvote-only system would look like, and I think it'd have to be like the opposite of StumbledUpon's 'like' button. You can only get at the downvote button on the page (hey, maybe only on the chapter page),[1] and the downvote button basicaly says "this is crap". So instead of likes/dislikes, we have something more like "6,000 people looked at this, and 538 thought it was trash" (or a simpler "15% disliked this" bar, or something. I dunno). Then you can maybe have the favorites button, which is supposed to be something else entirely[2], for stuff you really love or would recommend. Maybe call it the recommend button instead. Obviously, the 'recommend things for me' feature would give added weight to things in the recommend list of people you mark for such (like 814459's "favorites of followed" system). Of course, I'm not sure how much this discussion (which is oriented towards a FiMFic-like system) helps with what you're actually worried about.
[1] This is to
preventreduce downvote-swarming. Twilight, for example, would tend to be swamped with downvotes under a system like this if you could walk past it on the library shelf and hit 'downvote' rather than having to stop, pick it up, open the cover, and hit 'downvote'.[2] I would not be at all surprised if in practice it's not, to be honest. I hit 'favorite' on a rather high percentage of stories I read, separate from whether I'd recommend it. This is why I think 'favorites' and 'tracking' should be separate systems.
You could also have a slightly more in-depth process with more of a middle ground. On the far left is "loved it", just left of the middle ground is "pretty good" (note that there's a wide space between those), just to the right is "didn't like it", and all the way on the right is "hated it". This would not be on the same page as the work/whatever itself, you'd have to click through specifically to rate it. The (unlisted) default is 'meh, it was good enough'.
Obviously (well, quite evidently from reading this), a combination of reviews, network effects, and word-of-mouth is the ideal way to go. But that doesn't necessarily help either.
813032 You can do that currently with the
advanced searchfilters. It's a bit more work, but it can be done. I can talk you through it if you want.825386
Except that's called a star rating system, and we already had that. And it didn't work.
The simpler the system is, the less likely people are to find ways to misuse it. I'd be personally happy if downvoting came with a little box you had to type something into to explain why, but I know that'd rarely lead to any actual feedback if there wasn't any incentive to actually leave good critiques. To be honest, the system we have now that only gives the options to upvote, downvote, and/or favorite is probably the best we could get.
825431 No, not really. Lemme try and ilustrate.
Your 'input scale' looks something like this:
[]---------[]---|---[]---------[]
Those boxes are what's known as 'radio buttons'. That means you have to pick one. I listed the labels above, or we could reverse them.
The point is, you only see this if you cared enough to go seek it out and weigh in. Otherwise, you default to the middle position ("meh, whatever"). I'll grant this still over-represents the ends, but it does give us something between current-downvote's "this utterly sucks" and semidefault current-upvote.
The favorites system as currently implemented is worthless. It merges "stories-I-like-and-would-actually-recommend" (IE, favorites) with "stories-I-want-to-keep-track-of" (IE, web bookmarks, formerly available as "tracking"). Yes, in theory I could use Read-It-Later for tracking, but that's for stories I want to think about actually, you know, reading; and is overflowing enough as it is already.
825386 825431 How many stars were there, and how well did it work?
Netflix has a 5-star system, and it works pretty well. The difference is probably that Netflix users don't expect the people who made the movie will see their rating and feel sad about it. In practice a 5-star system is a 3-star system. People almost never use 1 or 2 stars, as you can see in the Netflix competition data.
How about "bad", "meh", and "good"? I've often wanted a "meh" button.
825613
Yes, the decision to combine "favorite" with "track" was bad in every way. It was supposed to simplify the interface, but I couldn't find the "track" button for weeks after I joined, because there wasn't one! It does not simplify it; it makes it more complicated.
826772
There is no 'meh' button. 'Meh' is "I didn't bother to vote".
What are the use-cases for a 'meh' button? 'Meh' is "I explicitly do not want to vote", or "I wish to change my vote from 'like'/'dislike' back to 'meh'" (which can be implemented with removing a vote). It's of much more use with something like here, where the voting buttons are readily available, than with something like what I was describing, where you have to actually seek out the voting buttons.
Re: Netflix star ratings: Probably what it comes down to is that people don't like to say something is bad unless it's really, utterly trash. So the bottom end (stars 1 and 2 on netflix, downvote here) get under-represented. (This would be the benefit of a downvote-only system: you implicitly upvote by showing up, and you can explicitly downvote if you want. Downvotes, being rarer on average, are weighted more than the implicit upvotes. This even better-represents traditional markets, where you can participate, write a glowing review and direct those who trust you to it, or write a bad review and direct people away from it.)
What I'm really interested in, at least from a FiMFiction perspective not necessarily from a general perspective, is views vs. repeat views. If I go back and read, say, Eternal once a month, evidently I really liked it and/or it had a huge impact on me. Should my upvote/favorite be weighted more because of that? What if I 'disliked' it, but still read it once a month?
Here's a really interesting case: dislike/favorite. What do we do with that?
(Example: I loved Steinbeck's Of Mice and Men (like/favorite). My brother strongly didn't like it, because of the ending (dislike). Consider a hypothetical someone who liked the writing, and enjoys/recommends it as a work of literature, but didn't like the ending (dislike/favorite).)
Tell me about it
. I did the exact same thing when I came here.
I actually don't mind having tracking but not favorites, I just wish it were called tracking. Literally everything else about the existing interface can stay the same, I just want it called tracking. That's a good idea, actually-- I'm going to go email knighty about it.
814326 It's funny how different things are better served by different systems. Comments, for example, work pretty well under an 'agree'/'disagree' system. Content, on the other hand, would like a bit more detail.
Not to knock Facebook's comment system, they've got a pretty decent-sounding thing going where you can agree cheaply (hitting 'like') but disagreeing is expensive (you have to actually comment and say that you disagree), which encourages explaining why you disagree right off. But that's comment systems.
PoweredByTea referenced this blog post and thus brought me here.
I've noticed something which seems relevant to this blog post. The highest-rated stories on FiMFiction are downvote magnets. Things which end up in the top attract downvotes at a wholly unusual pace. The spot is 'cursed' indeed, as was suggested here 823708
In order to stably remain on the top-rated page, stories have to get enough upvotes to survive arbitrary downvotes. Look at the upvote graph on Sunny Skies All Day Long. Hit that 'stats' button. It keeps drawing in viewers.
You know what? People aren't downvoting enough. I would be surprised if all of Sunny Skies' downvotes are legitimate. I bet that at least twenty of them are due to people kicking the story just for being on top of the list. It's not that the story couldn't have legitimately accumulated all of its (as of this writing) 77 downvotes, it's that the rate of voting on this site is so much lower than the rate of views. People who like stories don't upvote reliably. People who don't like stories don't downvote reliably.
When something new gets anywhere near the top of the top-rated list, it gets downvotes until it gets kicked off again. I keep watching this happen. I think things can only 'stabilize' on the top of that list when the trolldowns are all used up, and have all applied their singular downvotes. Trolldowns, in turn, paradoxically serve to keep the top rating list from having any kind of turnover. Old stories with standing momentum aren't threatened by such behaviors. It instead ensures that only things which have steamroller momentum from other sources can get onto the list.
I think most people think it's somehow 'wrong' to give downvotes. I frown every time I see comments on stories to the effect of 'I don't like it, but I'm not going to downvote'. That kind of behavior just leads to downvoting patterns that don't reflect story quality at all.
976722 I'm really sorry, but... I up-voted your comment.