Flash insight on voting chicanery

[SmilingLez]

Someday you are going go make an interesting point, instead of just shitting on everyone from some self anointed position of authority.
Today is not that day.
You seem to think there is some binary choice here.
Attack the process or defend the process.
Those aren't the only choices.
Most of us are addressing the very specific concerns raised by @LargoKitt. Such as the belief that someone with a large group of followers has a big advantage in the contests (leaving aside the obvious idea that one builds a large following by being a good writer, and thus more likely to win a contest).
Pointing out these specific flaws in Largo's argument is not the same as blindly defending the contest process.
As I mentioned earlier, many of us would likely do things differently. No one thinks the process is perfect. I'd be willing to bet Laurel doesn't think it's perfect. But it accomplishes it's intended purpose, and that's good enough.
I'm sure you've developed a perfect system, and we'd all love to hear about it...

An almost fair system would start with data instead of assumptions.

Take each Literotica category and analyze its number of entries, total views, ratings, average ratings, and positive and negative comments. Look at how much engagement stories in each category typically receive, and how that engagement differs between popular and less popular categories.

Then build a scoring system that accounts for those differences. A story in a category with a large built-in audience shouldn't automatically have an advantage over a story in a category that attracts far fewer readers. Lesbian fiction, for example, might naturally draw more traffic than non-erotic fiction. A fair system would need to account for that disparity rather than treat raw numbers as directly comparable.

Ratings could be normalized against each category's typical results, with safeguards against small sample sizes and unusually high or low votes. Views and comments could help measure engagement, but they shouldn't automatically count as proof of literary quality. Negative comments would need to be distinguished from constructive criticism, and comments and views could be manipulated too.

Would that produce a perfect system? No. No system is immune to bias or gaming. But it would be a data-driven starting point for developing a more equitable contest, rather than assuming the existing process is fair because it accomplishes its intended purpose.

You said you'd love to hear a better system. This is where I'd start: examine the data, identify the disparities, and design the rules around what the evidence shows.

And implementing a system like this would require considerably more than front-end cosmetic fixes. It would mean examining the existing database, analyzing the underlying data, developing and testing the scoring algorithms, and integrating them into the contest and voting workflows without breaking other site functions. That is backend engineering and software quality assurance, not a fresh coat of paint on the interface. The difficult work (took me 3 months to complete on TLC) is making the system function correctly, consistently, and transparently behind the scenes.

 

[Voboy]

Take each Literotica category and analyze its number of entries, total views, ratings, average ratings, and positive and negative comments. Look at how much engagement stories in each category typically receive, and how that engagement differs between popular and less popular categories.

Then build a scoring system that accounts for those differences.

You're proposing a GREAT deal of work for a pair of people who are already demonstrably backlogged.

The rating system, which led to the contest-scoring system, grew up as Literotica grew up. It has never changed. It's probable that it should have been continuously updated over time, but the site owners chose not to do so for whatever reason.

This happened over the past three decades. It works for the owners and it works for the readers, and it can be made to work for the writers. Though that last bit is less important than the other two, frankly.

At this point? The system is not in a position to be "built" based on... well, anything. Except that there's already an existing set of procedures that seems to work well enough for the site owners, who after all are the ones that matter here.

As many of us have pointed out, any change suggested wouldn't need to show that it's "better than" the existing system: it almost certainly would be, just by default, because it wouldn't be based on software processes that came of age in the late '90s. The suggested changes need to show that there's something about them that would make life easier for the site owners.

Define "easier" as... cheaper? Less labor-intensive? "Fairer?" However you wish. In almost no situation would an extensive new software upgrade, solely for the sake of "making voting fairer," be cheaper or less labor-intensive for the owners. Quite the contrary, actually. So that's why it's unlikely to happen: not because I think the status quo is best, or needs to be defended by me or anyone else, but because I'm being realistic.

Your assumptions about our motives aside, that's the best I can tell you.

 

[RejectReality]

An almost fair system would start with data instead of assumptions.

Take each Literotica category and analyze its number of entries, total views, ratings, average ratings, and positive and negative comments. Look at how much engagement stories in each category typically receive, and how that engagement differs between popular and less popular categories.

Then build a scoring system that accounts for those differences. A story in a category with a large built-in audience shouldn't automatically have an advantage over a story in a category that attracts far fewer readers. Lesbian fiction, for example, might naturally draw more traffic than non-erotic fiction. A fair system would need to account for that disparity rather than treat raw numbers as directly comparable.

Ratings could be normalized against each category's typical results, with safeguards against small sample sizes and unusually high or low votes. Views and comments could help measure engagement, but they shouldn't automatically count as proof of literary quality. Negative comments would need to be distinguished from constructive criticism, and comments and views could be manipulated too.

Would that produce a perfect system? No. No system is immune to bias or gaming. But it would be a data-driven starting point for developing a more equitable contest, rather than assuming the existing process is fair because it accomplishes its intended purpose.

You said you'd love to hear a better system. This is where I'd start: examine the data, identify the disparities, and design the rules around what the evidence shows.

And implementing a system like this would require considerably more than front-end cosmetic fixes. It would mean examining the existing database, analyzing the underlying data, developing and testing the scoring algorithms, and integrating them into the contest and voting workflows without breaking other site functions. That is backend engineering and software quality assurance, not a fresh coat of paint on the interface. The difficult work is making the system function correctly, consistently, and transparently behind the scenes.

Yeah, that kind of Voodoo math isn't going to cause any confusion or people crying foul. LOL

"Wait a minute? I had a 4.94. Why did this story with 4.89 beat me?"

"Well, it's in this category, which typical scores lower than yours, so it is algorithmically adjusted to account for that out of fairness."

"But, I beat a story in the same category with a 4.90."

"Well, this story has more votes than either of them, so there's another adjustment for that. Oh, and let's not forget the vote to view ratio, and..."

"So, you just pick who's going to win and make up some math to validate it then?"

 

[SmilingLez]

You're proposing a GREAT deal of work for a pair of people who are already demonstrably backlogged.

The rating system, which led to the contest-scoring system, grew up as Literotica grew up. It has never changed. It's probable that it should have been continuously updated over time, but the site owners chose not to do so for whatever reason.

This happened over the past three decades. It works for the owners and it works for the readers, and it can be made to work for the writers. Though that last bit is less important than the other two, frankly.

At this point? The system is not in a position to be "built" based on... well, anything. Except that there's already an existing set of procedures that seems to work well enough for the site owners, who after all are the ones that matter here.

As many of us have pointed out, any change suggested wouldn't need to show that it's "better than" the existing system: it almost certainly would be, just by default, because it wouldn't be based on software processes that came of age in the late '90s. The suggested changes need to show that there's something about them that would make life easier for the site owners.

Define "easier" as... cheaper? Less labor-intensive? "Fairer?" However you wish. In almost no situation would an extensive new software upgrade, solely for the sake of "making voting fairer," be cheaper or less labor-intensive for the owners. Quite the contrary, actually. So that's why it's unlikely to happen: not because I think the status quo is best, or needs to be defended by me or anyone else, but because I'm being realistic.

Your assumptions about our motives aside, that's the best I can tell you.

Dude I just answered the woman's question (I'm sure you've developed a perfect system, and we'd all love to hear about it...) based on something I've already done. Didn't say it was a perfect system said it was fairer

 

[SmilingLez]

Yeah, that kind of Voodoo math isn't going to cause any confusion or people crying foul. LOL

"Wait a minute? I had a 4.94. Why did this story with 4.89 beat me?"

"Well, it's in this category, which typical scores lower than yours, so it is algorithmically adjusted to account for that out of fairness."

"But, I beat a story in the same category with a 4.90."

"Well, this story has more votes than either of them, so there's another adjustment for that. Oh, and let's not forget the vote to view ratio, and..."

"So, you just pick who's going to win and make up some math to validate it then?"

Dude, you’ve just demonstrated why the methodology would need to be transparent and tested, not why it’s impossible. You’ve invented a deliberately convoluted algorithm, assumed it would be manipulated, and then argued against the scenario you made up.

I said analyze the data, account for differences between categories, and develop a fairer system. I didn't propose a secret formula that lets administrators pick winners after the fact.

If the data showed that category normalization made results less reliable, that would be a reason to revise the model. If it improved comparability, that would be a reason to use it. That's why you test a system against historical results before putting it into production.

And no, a 4.94 rating doesn't automatically prove a story is better than one rated 4.89. How many people voted, which category the story belongs to, and how representative those votes are all matter. Raw averages can look precise while concealing very different levels of support.

Your imaginary dialogue is amusing, but it doesn't refute the proposal. It just assumes that any result you don't agree with must have been manufactured.

 

[Voboy]

Dude I just answered the woman's question (I'm sure you've developed a perfect system, and we'd all love to hear about it...) based on something I've already done. Didn't say it was a perfect system said it was fairer

Excellent!

Now you can start your own erotica site, where I'm sure everything will run a lot better.

 

[SmilingLez]

Excellent!

Now you can start your own erotica site, where I'm sure everything will run a lot better.

Already did

 

[RejectReality]

Dude, you’ve just demonstrated why the methodology would need to be transparent and tested, not why it’s impossible. You’ve invented a deliberately convoluted algorithm, assumed it would be manipulated, and then argued against the scenario you made up.

I said analyze the data, account for differences between categories, and develop a fairer system. I didn't propose a secret formula that lets administrators pick winners after the fact.

If the data showed that category normalization made results less reliable, that would be a reason to revise the model. If it improved comparability, that would be a reason to use it. That's why you test a system against historical results before putting it into production.

And no, a 4.94 rating doesn't automatically prove a story is better than one rated 4.89. How many people voted, which category the story belongs to, and how representative those votes are all matter. Raw averages can look precise while concealing very different levels of support.

Your imaginary dialogue is amusing, but it doesn't refute the proposal. It just assumes that any result you don't agree with must have been manufactured.

You could post the exact mathematical formulas for everything in double sized font anywhere a mention of the contest exists, and you're going to get the exact same reaction from people that I just posted.

Transparency isn't enough. It's the ability of people to understand it ( often with little more than a glance ) that's going to determine how they react to it, and not a small number of people in this world treat anything beyond a simple average as witchcraft. When there's witchcraft, they're going to scream to burn the witch.

So, by your own metric of nobody needing to defend it, it fails. If you've actually implemented this sort of thing somewhere else, you'll find out soon enough — assuming it all doesn't collapse within a few months as most attempts to "do it better" typically do.

 

[Voboy]

Already did

Fantastic. Let us know how it goes.

On the rare occasions that you post here.

 

[SmilingLez]

You could post the exact mathematical formulas for everything in double sized font anywhere a mention of the contest exists, and you're going to get the exact same reaction from people that I just posted.

Transparency isn't enough. It's the ability of people to understand it ( often with little more than a glance ) that's going to determine how they react to it, and not a small number of people in this world treat anything beyond a simple average as witchcraft. When there's witchcraft, they're going to scream to burn the witch.

So, by your own metric of nobody needing to defend it, it fails. If you've actually implemented this sort of thing somewhere else, you'll find out soon enough — assuming it all doesn't collapse within a few months as most attempts to "do it better" typically do.

You're making a different argument now. Your first example portrayed category adjustments as arbitrary math used to manufacture a winner. Now you're arguing that even a legitimate, tested system might be too complicated for some people to understand.

That's a valid design consideration. If people can't understand the rules, the system needs a clearer explanation. It doesn't follow that the underlying methodology is unsound or that a simple average is inherently fairer.

And I never said nobody would question the results. I said a system should withstand scrutiny without requiring personal attacks on the people questioning it. Those are two very different things.

As for your prediction that attempts to improve existing systems usually collapse within months, that's a sweeping claim, not evidence against this particular proposal. You can't establish that a system would fail by assuming it will fail.

The point was to answer a sarcastic challenge with a concrete approach based on work I've already done. I wasn't promising a flawless system or claiming that every user would embrace it. I was proposing a starting point that could be tested, refined, and explained in terms people could understand.

Skepticism is reasonable. Treating anticipated resistance as proof that improvement isn't viable is just another assumption.

 

[SmilingLez]

Fantastic. Let us know how it goes.

On the rare occasions that you post here.

Don't worry, you'll hear about it when there's something worth saying. I prefer building things to spending years explaining why they can't be built.

 

[RejectReality]

You're making a different argument now. Your first example portrayed category adjustments as arbitrary math used to manufacture a winner. Now you're arguing that even a legitimate, tested system might be too complicated for some people to understand.

That's a valid design consideration. If people can't understand the rules, the system needs a clearer explanation. It doesn't follow that the underlying methodology is unsound or that a simple average is inherently fairer.

And I never said nobody would question the results. I said a system should withstand scrutiny without requiring personal attacks on the people questioning it. Those are two very different things.

As for your prediction that attempts to improve existing systems usually collapse within months, that's a sweeping claim, not evidence against this particular proposal. You can't establish that a system would fail by assuming it will fail.

The point was to answer a sarcastic challenge with a concrete approach based on work I've already done. I wasn't promising a flawless system or claiming that every user would embrace it. I was proposing a starting point that could be tested, refined, and explained in terms people could understand.

Skepticism is reasonable. Treating anticipated resistance as proof that improvement isn't viable is just another assumption.

Keep dreaming your pipe dream.

I've been at this (writing erotica for public consumption ) for 20 years. I've been under every scoring regime and contest judging method under the sun. They all end up with their Don Quixote sooner rather than later, who then starts a little PM or secret chatroom whisper campaign that then blows up into a full-blown tinfoil hat and aliens built the pyramids conspiracy.

 

[AwkwardlySet]

I must admit, it's quite fun to look at people arguing with an AI. Maybe this is how the war between man and machine starts!

 

[SmilingLez]

Keep dreaming your pipe dream.

I've been at this (writing erotica for public consumption ) for 20 years. I've been under every scoring regime and contest judging method under the sun. They all end up with their Don Quixote sooner rather than later, who then starts a little PM or secret chatroom whisper campaign that then blows up into a full-blown tinfoil hat and aliens built the pyramids conspiracy.

Twenty years of watching people argue about scoring systems, and you’ve apparently concluded that the safest way to evaluate a new idea is to ridicule it before it’s tested. That’s not experience proving a system can’t improve. It’s experience becoming a substitute for an argument.

 

[SmilingLez]

I must admit, it's quite fun to look at people arguing with an AI. Maybe this is how the war between man and machine starts!

Don't flatter yourself. There's a human behind this keyboard. I just happen to know how to construct an argument.

 

[Kelliezgirl]

An almost fair system would start with data instead of assumptions.

Take each Literotica category and analyze its number of entries, total views, ratings, average ratings, and positive and negative comments. Look at how much engagement stories in each category typically receive, and how that engagement differs between popular and less popular categories.

Then build a scoring system that accounts for those differences. A story in a category with a large built-in audience shouldn't automatically have an advantage over a story in a category that attracts far fewer readers. Lesbian fiction, for example, might naturally draw more traffic than non-erotic fiction. A fair system would need to account for that disparity rather than treat raw numbers as directly comparable.

Ratings could be normalized against each category's typical results, with safeguards against small sample sizes and unusually high or low votes. Views and comments could help measure engagement, but they shouldn't automatically count as proof of literary quality. Negative comments would need to be distinguished from constructive criticism, and comments and views could be manipulated too.

Would that produce a perfect system? No. No system is immune to bias or gaming. But it would be a data-driven starting point for developing a more equitable contest, rather than assuming the existing process is fair because it accomplishes its intended purpose.

You said you'd love to hear a better system. This is where I'd start: examine the data, identify the disparities, and design the rules around what the evidence shows.

And implementing a system like this would require considerably more than front-end cosmetic fixes. It would mean examining the existing database, analyzing the underlying data, developing and testing the scoring algorithms, and integrating them into the contest and voting workflows without breaking other site functions. That is backend engineering and software quality assurance, not a fresh coat of paint on the interface. The difficult work (took me 3 months to complete on TLC) is making the system function correctly, consistently, and transparently behind the scenes.

The difference between the highest rated story in LW and the highest rated story in Lesbian is .07 points. It's certainly possible to write a high scoring LW story.
The difference in average scores between the two is probably much higher. So writing a story to the formula in LW and getting the boost from normalized scores becomes the path to win or "game the system".
Sure you can come up with some Graded Respone Model, perhaps delve in Bayesian IRT if you are feeling really froggy.
But all that talk of "adjusting" things means people will see it no differently than here. If people are going to accuse Laurel of running sweeps until she gets the results she wants, why wouldn't they think the same of you?
Transparency?


There's a Graded Response Model Equation. What percentage of the AH can solve the equation? I'd venture not many. Then you'll have a self anointed math expert proclaiming your system is flawed, and you will have done a ton of work and not be any better off.

 

[AwkwardlySet]

Don't flatter yourself. There's a human behind this keyboard. I just happen to know how to construct an argument.

You definitely know how to construct a prompt. And you know well enough where I usually stand in these debates, so this isn't me defending Laurel's fan club; I'm just fed up with reading semi-senseless AI-generated debating points.

 

[RejectReality]

Twenty years of watching people argue about scoring systems, and you’ve apparently concluded that the safest way to evaluate a new idea is to ridicule it before it’s tested. That’s not experience proving a system can’t improve. It’s experience becoming a substitute for an argument.

No, it's having seen something fairly similar cause almost instant indignation and a torch-wielding mob. The site crashed and burned a few months later.

You're not treading any new ground.

 

[SmilingLez]

You definitely know how to construct a prompt. And you know well enough where I usually stand in these debates, so this isn't me defending Laurel's fan club; I'm just fed up with reading semi-senseless AI-generated debating points.

Three days in TLC's pending queue was apparently too much for you. Funny how you have so much patience for your own opinions but so little for a story review.

 

[Kelliezgirl]

Three days in TLC's pending queue was apparently too much for you. Funny how you have so much patience for your own opinions but so little for a story review.

Seems like just a few posts ago you were criticizing people for resorting to personal attacks...

 

[SmilingLez]

Seems like just a few posts ago you were criticizing people for resorting to personal attacks...

I stated a fact. He left his story on TLC in pending for 3 days before removing it. He's still listed as an active member. Don't believe me you can check for yourself

 

[Kelliezgirl]

I stated a fact. He left his story on TLC and pending for 3 days before removing it. He's still listed as an active member. Don't believe me you can check for yourself

Perfectly understandable why they'd have such a long que with all those stories going up.

 

[AwkwardlySet]

Three days in TLC's pending queue was apparently too much for you. Funny how you have so much patience for your own opinions but so little for a story review.

Oh wow, you are getting personal now. And quite unfair.

I wrote what I intended to be an exclusive story for your website, all with a wish to help out and give my contribution to a fledgling story site. I submitted a story and waited for over three days. Zero response, zero acknowledgement of any kind. The site was pretty dead during that period, if I remember correctly, in the sense that no new stories were published, so you weren't exactly swamped with material.

Now, if I'm not mistaken, you started that website to do things differently compared to Lit, yet it turned out I was getting the Lit treatment all over again. So yeah, I got annoyed and canceled the submission. And that's fully on you as an admin. I mean, if you aren't taking your site seriously, why should I? Once again, I made all the effort of writing a story for your site exclusively, just to help with content. Rubbing this in my face now is more than unfair.

 

[Voboy]

Oh wow, you are getting personal now. And quite unfair.

I wrote what I intended to be an exclusive story for your website, all with a wish to help out and give my contribution to a fledgling story site. I submitted a story and waited for over three days. Zero response, zero acknowledgement of any kind. The site was pretty dead during that period, if I remember correctly, in the sense that no new stories were published, so you weren't exactly swamped with material.

Now, if I'm not mistaken, you started that website to do things differently compared to Lit, yet it turned out I was getting the Lit treatment all over again. So yeah, I got annoyed and canceled the submission. And that's fully on you as an admin. I mean, if you aren't taking your site seriously, why should I? Once again, I made all the effort of writing a story for your site exclusively, just to help with content. Rubbing this in my face now is more than unfair.

Shhhhhhh.

I'm told that site works really well. I heard it from a great source somewhere, recently.

 

[Kelliezgirl]

Shhhhhhh.

I'm told that site works really well. I heard it from a great source somewhere, recently.

The contests are amazeballs!