What is a good rating?

[EarlyMorningLight]

A better option might be to use the "trimmed mean" like they do in the olympics, where you would leave out a certain percentage of the highest and lowest scores and take the average of the rest.

That could work too. I guess it’s just my brain trying to solve the problem of “if a story has 22 five votes and 4 one votes, is 4.38 a useful representation of how most readers felt about the story?”

 

[TheRedChamber]

As a new writer here, I’m watching my second story hover just under 4.5. I’d love to get the red H just because I’m still trying to build a following, and the visibility would help. That being said, I know my work is a bit difficult. I’m not trying to color inside the lines. Both of my stories so far have content that crosses categories, there’s complex emotional content, and I’m not writing HEA stories. By Lit standards, I’m asking a great deal of the readers, so I’m pretty damn happy with how they’re performing.

I'm in a similar position with my latest story (thanks for your kind comment on that one BTW) - a sweep might push it over but it looks increasingly like it's going to stablize at about 4.4. That's a bit disappointing as I was particularly happy with it and especially as my latest obsession is trying to get to at least 50% of my stories rated H, but there's a learning experience with the various catagories and readerships on Lit and I'm increasingly looking at the stories thinking 'If I do/don't do this, it'll probably score higher, but do I want to?'

I'm kind of interested in the score for your story because both of my BDSM stories got H's while not necessarily being any better in my mind than other stories that didn't. Both of them are light-hearted and gentle on the pain and I was initially surprised that they did so well. Your story is clearly harsher on the sex and more emotionally complex, but has done well even if it hasn't quite hit H yet. BDSM reader do seem to be more likely to vote for intelligently written work from what I can make out.

That could work too. I guess it’s just my brain trying to solve the problem of “if a story has 22 five votes and 4 one votes, is 4.38 a useful representation of how most readers felt about the story?”

I don't think it's a problem that is going to be solved by mathematics. Most people who vote tend to vote 5 (doesn't mean our stories are great, only that people who really liked it are going to vote) and there may be a few people who dislike the story enough to vote with a 1. More difficult, 'controversial' or atypical stories are always going to pick up a few more of the 1s. It might be more straightforward to go with a simple binary 'like','dislike' system because there are tons of stories which I'd give a 4 to but I'd feel bad about lowing the score which is usually already above 4. I think you have to come to terms with the fact that people are going to dislike stories for reasons which are not always directly connected with the quality.

 

[EarlyMorningLight]

I'm kind of interested in the score for your story because both of my BDSM stories got H's while not necessarily being any better in my mind than other stories that didn't. Both of them are light-hearted and gentle on the pain and I was initially surprised that they did so well. Your story is clearly harsher on the sex and more emotionally complex, but has done well even if it hasn't quite hit H yet. BDSM reader do seem to be more likely to vote for intelligently written work from what I can make out.

I knew going in that combining a partner-sharing gangbang with harsh BDSM and the emotional minefield of a relationship unraveling would likely get me some downvotes for content, even with meticulous disclaimers upfront. It’s the nature of the beast. I find it strange that so many people on a smut site get prudish about things outside of their proclivities.

I was actually talking about this with rvagirl (weird, tagging isn’t working on my phone), whose recent story in I/T copped hate from at least one user for having supporting characters who were LGBTQ+, even though they were nowhere near the sexual activity. The cognitive dissonance of getting offended by a lesbian and a trans character when you’re reading a story about brother-sister fucking is both mind-blowing and hilarious to me.

 

[ElectricBlue]

The cognitive dissonance of getting offended by a lesbian and a trans character when you’re reading a story about brother-sister fucking is both mind-blowing and hilarious to me.

That's Lit for ya .

 

[Bramblethorn]

The median wouldn't work so well in practice. If you still allowed people to vote 1-5 stars, then the median score would almost always be a whole number. For example, all the stories that currently have a red H (and many lower-scoring stories as well) would get a median score of 5, and there wouldn't be any way to distinguish between them.

A better option might be to use the "trimmed mean" like they do in the olympics, where you would leave out a certain percentage of the highest and lowest scores and take the average of the rest.

Some thoughts on that approach:

#1: For stories that are currently at the very high end of the range, you lose a bit of discrimination. For the sake of discussion, let's assume we ignore the top and bottom 10% of votes - then everything currently scoring 4.90 or higher would become perfect 5.0.

I don't see this as a big problem. It might be a bit of a nuisance for contests and toplists creating ties at the top, but that's already an issue with rounding scores to two decimal places and we seem to manage. For readers, I'm not sure there are many people who are going to think "I'll read a 4.95 but not a 4.90"; both of those are very high scores and if somebody's choosing between them, it'll probably come down to the concept or the author rather than the exact number.

#2: You lose some ability to discriminate between love/like and love/hate. A story that scores 90 x 5 and 10 x 4 will end up with the same trimmed mean as one that gets 90 x 5 and 10 x 1.

If one believes that all 1s are bad-faith votes, that's not a huge problem. One might even argue that the second story should be rated higher because the 4s are likely to be genuine but the 1s aren't.

OTOH... some stories really are polarising. If you applied Literotica-style voting to something like "Atlas Shrugged" you'd probably get a bunch of fives and a bunch of ones with not much in between, and both the 5s and the 1s would be mostly sincere votes. The difference between "10% of readers liked but didn't love it" and "10% of readers thought it was awful" is probably worth knowing.

(FWIW, I don't think any way of reducing voting to a single number is going to handle this well; in the end, we're trying to collapse multi-dimensional information into a single dimension and some kind of information loss is inevitable.)

#3: Except for stories at the very high end of the range, this may not be as effective at nerfing trolls as one would hope.

Suppose I have an okay-but-not-great story up with 100 votes. Without trolling, the scores might look something like this:
50 x 5 (or after trimming the top 10%, 40 x 5)
30 x 4
15 x 3 (or after trimming the bottom 10%, 10 x 3)
5 x 2 (after trimming 0 x 2)

Raw average: 4.25. Trimmed mean (ignore top and bottom 10%): 4.38.

Now some asshole comes along and drops ten one-bombs on it:

50 x 5 (after trimming, 39 x 5)
30 x 4
15 x 3
5 x 2 (after trimming, 4 x 2)
10 x 1 (after trimming, 0 x 1)

Raw average: 3.95 (drop of 0.3). Trimmed mean: 4.18 (drop of 0.2).

Even though trimming excludes ALL the one-stars, it means we now have five three-star and four two-star votes being included that previously wouldn't have been. It softens the impact a little but the trolling still makes a substantial difference to the score.

 

[Bramblethorn]

There's no panacea for malicious people, etc. But it seems to me that the ideal (which of course will never happen) would be to try to drive the system further toward helping people find stuff they like, and away from the "artificial stack-ranking of goodness" model.

Like, a really basic issue is that a story can end up "3.8" rated for all kinds of reasons - it could have sloppy writing, it could have a boring/unoriginal premise, it could have a kink that turns off a good number of people, it could have some idiosyncratic quirk that makes a few people mad enough to drop 1' on it, etc. etc. It's all glommed together. But for readers, the details here may matter a lot. Some people don't care about writing mechanics if the characters are great. Some people don't care if it's unoriginal as long as it's a good stroke. So, as far as helping decide whether a story is worth reading for YOU, "3.8" appears to me to be pretty close to meaningless. Yet, I believe a great many readers tend to think of it as determinative.

Hard agree; I'd also note that most of us here default to an author's perspective on scoring, which may not match readers' perspectives or priorities.

Assigning a single number to each story is always going to be an extremely imperfect tool. Rather than trying to make it slightly less imperfect, it'd be better to ditch that mentality altogether and look for better solutions.

Literotica's category system is a late-1990s-tech attempt at a solution to that problem. The "similar stories" function is another fairly low-tech approach. In 2022, there are many more sophisticated options available - not simple to implement and none of them foolproof, but some probably a good deal more useful to readers than a single score.

(In fairness, I'd also note that many of those options depend on having large numbers of stories and readers to analyse; even if they'd been available in the early days, they might not have been terribly effective when Lit was a smaller site.)

I think the truth is that ANY story on Literotica would turn off more people than it turned on, if everyone read it.

This is true but it's bad manners to say it here in front of authors ;-)

 

[Sunadmire]

Firstly, an admission: I only read about half of the first page's comments to the threads before realizing that there is a lot more. Then I skipped to page 5 and read the latest comments.
As reader/writer, I wanted to add my penny's worth on observations.
1. I am trying to write good stories and appreciate ti when the stories get a lot of reads. Readers seems to appreciate my scribblings with 7 out of my 12 stories 'Hot' and 5 more hovering above 4.4. Even for these stories, the score dips notably when a 1-star comes along. I learn nothing from these low scores, and would appreciate it if the reader told me why the low score was given.

2. It seems as if Lit is picking up on the '1-bombers' as I notice a reduction in votes and an increase in the score - well done for Lit.

3. Someone on page 1 made the remark that scoring of the stories are supposed to reflect a bell curve. That would be true if a normal distribution is assumed. I would actually expect a Weibull distribution with 'k' (or gamma) = 2. Lots of high scoring stories because most authers, like me want to be proud of their story/stories. Unfortunately there seems to be authors producing crap in the hope that one of them may hit the mark.

4. It is heart breaking for an author not to receive feedback/appreciation for a story. A lot of effort is put into the creation thereoff. When I joined Lit, I only read 'Hot' stories and never scorred or commented. Since the publication of my first story, I started to read lower rated stories (3 5 to 4.5) and tried to add value in my score. If the story deserves a low score, I would score appropriately. A good story (with some issues like bad editing) would get a 4 and an exceptional story would get a 5.

5. As several others noted the hypocrisy of some readers. They are reading erotic stories and then wanted to be moral councelors on unwanted pregnancies and STD's. Since the stories are fictional, these readers should get past their issues and just enjoy the story. I started to list 'tags' at the beginning of the story, to advise readers on the expected content. It should both entice readers to read, and discourage readers not into objectionable behaviour from reading it. I have come accross stories submitted in a wrong category, which got penalised for it.

6. I have a sweet spot for stories that I read (3 000 to 10 000 words). Long enough to allow good character development, but short enough to have an easy/quick read. I wish Lit would add the word count of stories so I can choose a good one to read directly. I often have to open a story, only to see it has a word count of say 43 000 words. When I am in the mood for a long story, I would likewise be able to spot one without an extensive search.

7. The two stories with low scores were submitted for the 750 word challenge. I intended to write only one in which a white dude have sex with a black woman - currently at 4.01 from 143 votes with 10 likes form 17k reads. I immediately noted a polarization in voting, receiving a lot of 5's and a lot of 1's, with little in-between. I suspected malicious voting by people who didn't like white on black sex and wrote a second black on white to test my suspicion. I got mixed results with 3.74 from 137 votes and 16 likes from 14.5k reads. Comments requested longer stories (which I am working on). My conclusions are:
a. 750 words doesn't allow for good character development.
b. More likes doesn't equate to higher scores.
c. White on black sex and black on white sex have both their own niche in the erotic story space.
d. It was fun writing these stories, will do it again..

All for now

Sun

 

[Sunadmire]

If I can add something to my earlier post:

8. Two of my stories barely achieved more than 2000 reads while these were stil 'new'. A third managed just over 3000. This, in spite of scoring relatively high. I tracked the progress on the third of there low read stories and noticed that it slipped of the 'new' list on day 3, depriving it of the exposure before the week was over. This was due tobthe high number of stories submitted at the time. I hope Lit can increase the amount of srories on the new list to give authors the exposure they expect after putting in time and effort.

Best regards

Sun

 

[000zing]

A good rating is when you yourself are happy with what you've produced. Nothing else matters.