Score Vandalism

[EmilysRopes]

And that's all based on storytelling, not mechanics, isn't it? This isn't a judging question, just a check against a post on another board separating reader response between storytelling and mechanics. If you're saying what motivates you is the attractiveness of the storytelling, you are evidencing the other post's conclusion that the quality of storytelling is more important to the preponderance of readers here than that of mechanics.

Absolutely it is about storytelling though mechanics could mean the mechanics of the sex or the mechanics of the writing. Both matter a little, for me the sex should be believable and the stories should be competently written. But bothering to score is quite rare for me. The occasional 5 as a thank you, the even rarer 1 as a please be more humane.

But maybe why and how people score is a different subject.

 

[M_K_Babalon]

Any type of voting is pretty much exploitable. I suppose if voting locked once a story got a Red H, would help very few.

People are probably not gonna like this idea, but here's a theory; get rid of actual voting, do more like Wattpad or AO3 where it's just a like that you can't take back. Keep the Red H to be had after so many likes. Without the scale rating the degens can't neg rep. This might lead to fewer voting, but the likes will be sincere because either you like a story or you don't, there's no kinda or maybe involved.

it won't matter if somebody is anon or not, and it might, just might boost comments. Users anon or mot will look at a story and instead of seeing 3.64 score, they'll see 364 stars, which may get the story read, where so many would avoid it by the 3.64 rating. Given the chance of fewer votes because of it, the H won't exactly need high vote rate to get, or it could be around 200 stars, or be done away, although that's what many look for.

I know this will never happen, but it sounds better than most of these alternatives, and still benefits readers, because there's numbers. Maybe even make registered users likes are worth double. It'll show how many people liked the story more than the current system does.

Somebody said something about thumbs up&down, which might not be bad, given it works like youtube and not reddit. I hope all of this makes sense, I just got home from having several teeth pulled.

 

[AwkwardlySet]

I hope all of this makes sense, I just got home from having several teeth pulled.

You were interrogated by the mob?!

 

[M_K_Babalon]

You were interrogated by the mob?!

Nah, but it was torture.

 

[Bramblethorn]

You're better at the math than I am, but would this necessarily be true if you lopped an equal number of 5s? I don't know what the right percentage is, but if you lop off the highest votes and lowest votes in equal measure it would seem to me as a matter of common sense you'd get a score that would approximate what most people think minus the outliers.

If lopping the bottom 10% of scores leaves you with nothing but 5s (which is the case for anything with 4.9 or higher, and some below that score) then lopping the top 10% as well still leaves you with all 5s.

This wouldn't be true for all categories, either, because many categories have top scores significantly below 5.

Which ones are you thinking of?

I haven't checked through every category, but let's take LW as an example of a category famous for tough scoring and bombing. Even there, there are thirteen stories ranked 4.85 or higher in LW. So what happens to those stories if we trim the top and bottom 10%?

For the sake of discussion, I'm going to consider a story which has an average of 4.85 with exactly 100 votes; obviously the actual vote counts in LW will be higher, so multiply the following counts accordingly.

If we suppose that the only votes cast are 5s, 4s, and 1s, then the possibilities are:

85 x 5*, 15 x 4*, 0 x 1*: after trimming, this leaves us with 75 x 5* and 5 x 4*, for an average of 4.94.
88 x 5*, 11 x 4*, 1 x 1*: after trimming, this leaves us with 78 x 5* and 2 x 4*, for an average of 4.98.
91 x 5*, 07 x 4*, 2 x 1*: after trimming, this leaves us with 80 x 5* and nothing else, for an average of 5.00.
94 x 5*, 03 x 4*, 3 x 1*: as above.

OTOH, if we suppose that the bombers are sneaky and dropping 2s instead of 1s, we get these cases:

87 x 5*, 12 x 4*, 1 x 2*: after trimming, 77 x 5* and 3 x 4*, average 4.96.
89 x 5*, 09 x 4*, 2 x 2*: after trimming, 79 x 5* and 1 x 4*, average 4.99.
91 x 5*, 06 x 4*, 3 x 2*; or 93/3/4; or 95/0/5: after trimming, all 5*, average 5.00.

We could explore other situations with mixes of 1s, 2s, 3s, 4s, and 5s, but it boils down to: if we suppose those stories are getting a significant amount of bombing, then trimming the top and bottom 10% would push them all up to perfect 5s. If we suppose they're not getting a significant amount of bombing (like in the case where the votes are all 4 and 5s) then it won't necessarily push everything up to 5.00...but then if bombing is that low (in LW! on a toplist!) why are we doing this in the first place?

If we modify that to, say, just trimming the top and bottom 5%, the effects are milder. But then it becomes less effective for stopping genuine bombing, and it's still likely to increase the graininess of the toplists.

 

[Bramblethorn]

Nah, but it was torture.

Worse than another round of "here's how to fix Literotica scoring" though?

 

[Bamagan]

Worse than another round of "here's how to fix Literotica scoring" though?

 

[NoTalentHack]

If lopping the bottom 10% of scores leaves you with nothing but 5s (which is the case for anything with 4.9 or higher, and some below that score) then lopping the top 10% as well still leaves you with all 5s.



Which ones are you thinking of?

I haven't checked through every category, but let's take LW as an example of a category famous for tough scoring and bombing. Even there, there are thirteen stories ranked 4.85 or higher in LW. So what happens to those stories if we trim the top and bottom 10%?

For the sake of discussion, I'm going to consider a story which has an average of 4.85 with exactly 100 votes; obviously the actual vote counts in LW will be higher, so multiply the following counts accordingly.

If we suppose that the only votes cast are 5s, 4s, and 1s, then the possibilities are:

85 x 5*, 15 x 4*, 0 x 1*: after trimming, this leaves us with 75 x 5* and 5 x 4*, for an average of 4.94.
88 x 5*, 11 x 4*, 1 x 1*: after trimming, this leaves us with 78 x 5* and 2 x 4*, for an average of 4.98.
91 x 5*, 07 x 4*, 2 x 1*: after trimming, this leaves us with 80 x 5* and nothing else, for an average of 5.00.
94 x 5*, 03 x 4*, 3 x 1*: as above.

OTOH, if we suppose that the bombers are sneaky and dropping 2s instead of 1s, we get these cases:

87 x 5*, 12 x 4*, 1 x 2*: after trimming, 77 x 5* and 3 x 4*, average 4.96.
89 x 5*, 09 x 4*, 2 x 2*: after trimming, 79 x 5* and 1 x 4*, average 4.99.
91 x 5*, 06 x 4*, 3 x 2*; or 93/3/4; or 95/0/5: after trimming, all 5*, average 5.00.

We could explore other situations with mixes of 1s, 2s, 3s, 4s, and 5s, but it boils down to: if we suppose those stories are getting a significant amount of bombing, then trimming the top and bottom 10% would push them all up to perfect 5s. If we suppose they're not getting a significant amount of bombing (like in the case where the votes are all 4 and 5s) then it won't necessarily push everything up to 5.00...but then if bombing is that low (in LW! on a toplist!) why are we doing this in the first place?

If we modify that to, say, just trimming the top and bottom 5%, the effects are milder. But then it becomes less effective for stopping genuine bombing, and it's still likely to increase the graininess of the toplists.

 

[SimonDoom]

If lopping the bottom 10% of scores leaves you with nothing but 5s (which is the case for anything with 4.9 or higher, and some below that score) then lopping the top 10% as well still leaves you with all 5s.



Which ones are you thinking of?

I haven't checked through every category, but let's take LW as an example of a category famous for tough scoring and bombing. Even there, there are thirteen stories ranked 4.85 or higher in LW. So what happens to those stories if we trim the top and bottom 10%?

For the sake of discussion, I'm going to consider a story which has an average of 4.85 with exactly 100 votes; obviously the actual vote counts in LW will be higher, so multiply the following counts accordingly.

If we suppose that the only votes cast are 5s, 4s, and 1s, then the possibilities are:

85 x 5*, 15 x 4*, 0 x 1*: after trimming, this leaves us with 75 x 5* and 5 x 4*, for an average of 4.94.
88 x 5*, 11 x 4*, 1 x 1*: after trimming, this leaves us with 78 x 5* and 2 x 4*, for an average of 4.98.
91 x 5*, 07 x 4*, 2 x 1*: after trimming, this leaves us with 80 x 5* and nothing else, for an average of 5.00.
94 x 5*, 03 x 4*, 3 x 1*: as above.

OTOH, if we suppose that the bombers are sneaky and dropping 2s instead of 1s, we get these cases:

87 x 5*, 12 x 4*, 1 x 2*: after trimming, 77 x 5* and 3 x 4*, average 4.96.
89 x 5*, 09 x 4*, 2 x 2*: after trimming, 79 x 5* and 1 x 4*, average 4.99.
91 x 5*, 06 x 4*, 3 x 2*; or 93/3/4; or 95/0/5: after trimming, all 5*, average 5.00.

We could explore other situations with mixes of 1s, 2s, 3s, 4s, and 5s, but it boils down to: if we suppose those stories are getting a significant amount of bombing, then trimming the top and bottom 10% would push them all up to perfect 5s. If we suppose they're not getting a significant amount of bombing (like in the case where the votes are all 4 and 5s) then it won't necessarily push everything up to 5.00...but then if bombing is that low (in LW! on a toplist!) why are we doing this in the first place?

If we modify that to, say, just trimming the top and bottom 5%, the effects are milder. But then it becomes less effective for stopping genuine bombing, and it's still likely to increase the graininess of the toplists.

OK. You have me convinced, mostly. I thought perhaps we'd finally stumbled upon a way to change the system that would provide some improvements without the bad side effects, but it doesn't look like that's so.

 

[pink_silk_glove]

OK. You have me convinced, mostly. I thought perhaps we'd finally stumbled upon a way to change the system that would provide some improvements without the bad side effects, but it doesn't look like that's so.

Top and bottom 20%?

 

[M_K_Babalon]

Worse than another round of "here's how to fix Literotica scoring" though?

Five teeth, wouldn't let me wear my headphones, had me on an iv but not enough to knock me out. Hurt like fuck and I was pissed. Spent ten minutes mourning my teeth once I got home and the two ibuprofens and hydros I took started to kick in.

 

[Voboy]

Top and bottom 20%?

At a certain point, it's no different than the sweeps, in terms of overall effect.

 

[Bramblethorn]

OK. You have me convinced, mostly. I thought perhaps we'd finally stumbled upon a way to change the system that would provide some improvements without the bad side effects, but it doesn't look like that's so.

It's not that it's technically impossible. There are ways to reduce the impact of outlier scores without breaking stuff like toplist separation or creating weird effects like "one vote has no effect on scores/next vote has huge effect on scores" - you can define something that's partway between an average and a median, for instance.

It's more that anything which addresses those issues ends up being something that only a maths nerd could love, and while I resemble that description, I do think there's a lot to be said for having a system that's simple enough for most users to understand. Anything fancier than an average is really pushing it. Even averages are too complex for some.

 

[RejectReality]

There's no fixing scoring. Removing that accursed red H from the mix would do a lot, but that's no more likely than getting rid of anonymous. The eventual condensing of multi-part stories in the toplists will do much to improve those, as they won't be overwhelmed by multiple entries of the same story. ( That is already on the books per Laurel and Manu in earlier posts. They're likely testing the shit out of it because the toplists have proved problematic in other technical ways in the past )

10 points instead of 5? SOL does that and there's little difference in the score distribution or usefulness. ( The math-voodoo only gives the illusion of wider distribution ) Break the score down into multiple parts? SOL tried that and abandoned it because the overwhelming majority of people don't use it as intended, and instead put the same score across the board. Like only? Lush does that and all somebody has to do is distribute a few nudes that look to be the same person and everything else gets buried under the drooling simps slamming the like button with their dicks. Every security measure you attempt to put in place will get circumnavigated — likely within minutes of implementation. Every one will also inevitably make the user experience for at least some regular users a pain in the ass.

The only thing that kind of works is exactly what Lit does now. Policing the votes with sweeps after the fact.

The solution to a troll bombing campaign is to keep writing and building your following. The odds are that whatever you did to raise some troll's ire, somebody else is going to do something worse or at least more recent to draw away their attention eventually. ( LW and other Lit-controversial subjects notwithstanding. If cuckold, reconciliation, femdom, bi-male, etc. is your thing, you're just going to have to learn to live with it, and/or crosspost to other sites where those works aren't subject to such hatred. Cultivate multiple venues, folks. ) Those followers WILL eventually create a buffer that blunts the troll's efforts. There's a good chance that eventually ( sometimes even years later ) they'll make a mistake that will get their poo-flinging removed from your work as well.

 

[Cacatua_Galerita]

This isn't a critique site though. It's a porn stories site. No voter should have to justify their vote (low or high).

Reddit has posted how they think that users should use voting, and not a single person actually follows them, unless they are trying to shame people who downvote. And people only ever complain about downvotes, never about upvotes.

Stop trying to fix the voting system, because readers are going to vote how they vote. It's much more sane to acknowledge how the readership votes and move on with your life. You cannot enforce how people vote unless you want to drive down engagement.

As mentioned, with the vast bulk of the stories sitting somewhere between four- and five-stars, the influence of a one-star vote is out of proportion to the “typical” vote. The undue power of a one-bomb leaves it open to abuse.

Under my suggestion, if the reader is logged-in (i.e. not anonymous) then they can one-bomb at will. As always, they can only do it once per story and the site keeps a record of their voting history.

Only if a reader is anonymous – and is giving a one- or two-star vote – would they need to “justify their vote.” And if they can’t be bothered to at least write “boring” then they obviously didn’t “hate it” enough to be wielding the immense power of a one-bomb.

Compared to the effort authors put into publishing a story, it is trivially easy for someone to one-bomb it into oblivion (i.e. knock it out of the 30-day and 12-month “Top Lists” so it gets fewer views). If they want to do that much damage to a story’s prospects, then yes; justify that vote.