Weird rating goings-on

[Senor_Smut]

I don't usually check the ratings for my stories, but I happened to glance at the page today and saw something weird. A few weeks back, one of my stories had a rating of something like 4.38, and today the same story has a 4.57 rating on the same number of ratings as it had before. Has this happened to anyone else?

 

[EmilyMiller]

I don't usually check the ratings for my stories, but I happened to glance at the page today and saw something weird. A few weeks back, one of my stories had a rating of something like 4.38, and today the same story has a 4.57 rating on the same number of ratings as it had before. Has this happened to anyone else?

There was a sweep at the end of Summer Lovin’ - you had some votes deleted and some new ones replace them. It’s normal (sadly).

 

[Senor_Smut]

There was a sweep at the end of Summer Lovin’ - you had some votes deleted and some new ones replace them. It’s normal (sadly).

OK now I'm more confused than before. Why would votes be deleted?

 

[alohadave]

OK now I'm more confused than before. Why would votes be deleted?

A sweep removes votes that it considers to be fraudulent or suspicious.

We don't know how it works, and the site doesn't want us speculating publicly about how it works.

 

[Senor_Smut]

A sweep removes votes that it considers to be fraudulent or suspicious.

We don't know how it works, and the site doesn't want us speculating publicly about how it works.

Well I'll be damned (obviously, but I mean more so than usual). I don't know what to make of that. I'm glad I don't pay attention to the numbers or that would drive me bonkers.

 

[OralAllOver]

Inflation

 

[Five_Inch_Heels]

AI, submissions, voting, and now ratings, all within a few posts of each other.

 

[Actingup]

Well I'll be damned (obviously, but I mean more so than usual). I don't know what to make of that. I'm glad I don't pay attention to the numbers or that would drive me bonkers.

It's a good thing IMHO - the site owners care about vote manipulation, and try and deal with it. It's imperfect, but much better than not caring. There are authors here who can tell you with a reasonable degree of confidence exactly what bombs they've received and when, and now many votes have been 'swept'.

 

[EmilyMiller]

It's a good thing IMHO - the site owners care about vote manipulation

If they really cared, there are other things they could do, including the non-radical proposal I made to Laurel and Manu.

 

[Actingup]

If they really cared, there are other things they could do, including the non-radical proposal I made to Laurel and Manu.

I think it's a solid proposal, similar to gymnastics voting? And also, thank you for linking to that picture of the tight bear t-shirt.....

 

[feel_the_beast]

I too have had unusual [and I like to think undeserved] low ratings removed, periodically, every few months. Don't know how the process works, and since it was said they would rather we not speculate, I will refrain. But I too am glad it's being done. Authors of much longer experience than I have concurred that there are trolls here who run amok dropping '1' ratings on stories they have not even read just for the hell of it.

 

[EmilyMiller]

I too have had unusual [and I like to think undeserved] low ratings removed, periodically, every few months. Don't know how the process works, and since it was said they would rather we not speculate, I will refrain. But I too am glad it's being done. Authors of much longer experience than I have concurred that there are trolls here who run amok dropping '1' ratings on stories they have not even read just for the hell of it.

Having to have a batch process remove some (far from all) malicious votes is a tacit admission that the system is dysfunctional. Sweeps are a semi-effective BandAid which can be easily avoided with just a little effort. They aren’t a solution.

 

[RejectReality]

If they really cared, there are other things they could do, including the non-radical proposal I made to Laurel and Manu.

Wouldn't do anything to reduce contest manipulation. It's plainly obvious if you watch it long term the malicious voting happens until the competition is low enough to no longer threaten their darling. Reducing the impact of a troll or cheerleader vote would only mean more of the same would get cast until the desired result was achieved.

Granted, the additional malicious voting would increase the odds of the votes getting swept, because they'd inevitably leave more traces.

You simply cannot reduce the ability of bad actors to manipulate scores to any significant degree without negatively impacting the experience of the average user. When you make it harder for the average person to vote, they don't. At which point, every vote a bad actor casts goes up in value.

And it's math. Anything more complicated than a simple average will generate torch-wielding mobs of authors screaming "Witchcraft!"

 

[EmilyMiller]

Wouldn't do anything to reduce contest manipulation.

I disagree.

It's plainly obvious if you watch it long term the malicious voting happens until the competition is low enough to no longer threaten their darling. Reducing the impact of a troll or cheerleader vote would only mean more of the same would get cast until the desired result was achieved.

They would have to work a lot harder. And we need to drop the fan club fiction. Only actually authors care enough to game the system. We all know how low user engagement is.

Granted, the additional malicious voting would increase the odds of the votes getting swept, because they'd inevitably leave more traces.

They would have to vote many more times and it would stick out as a malicious pattern more obviously.

You simply cannot reduce the ability of bad actors to manipulate scores to any significant degree without negatively impacting the experience of the average user.

I disagree - my proposal does nothing to impact user experience.

When you make it harder for the average person to vote, they don't. At which point, every vote a bad actor casts goes up in value.

This is a red herring, it’s not what I’m proposing.

Nothing in my proposal makes it harder for bona fide people to vote.

It’s exactly the same to vote. The central premise of my suggestion is totally mainstream and at the heart of most rating systems thaf attempt to make gaming harder.

And it's math. Anything more complicated than a simple average will generate torch-wielding mobs of authors screaming "Witchcraft!"

This is pure opinion. Similar approaches are adopted across the Internet without causing riots.

 

[Senor_Smut]

Weird. I mean, I wasn't even in the competition, or if I was, I didn't know about it. I do know there are a few folks who have decided they really don't like me and 1-star everything I put up, because I ignored their recommendations for how I write my stories. It's all very strange -- so much ego around posting free smutty stories on the internet! It's not like any of us makes more money for a 5-star story than for a 1-star story -- 5 x 0 = 0. I guess some people really have their self-image wrapped up in this.

 

[iwatchus]

I think some people just take glee in making other people unhappy, sad as that is. I had a bunch of stories above 4.9. Someone went through yesterday or the day before and hit almost all of them with a low score (2's mostly I think), all in the same hour or two. (Yes I check my scores too often.)

 

[EmilyMiller]

I wasn't even in the competition

The sweeps encompass all stories, not just the comp ones. I’ve had like fifteen of my stories leap from 4.4x to 4.6x in previous sweeps.

 

[Senor_Smut]

The sweeps encompass all stories, not just the comp ones. I’ve had like fifteen of my stories leap from 4.4x to 4.6x in previous sweeps.

Huh. Learned something today.

 

[RejectReality]

I disagree.

They would have to work a lot harder. And we need to drop the fan club fiction. Only actually authors care enough to game the system. We all know how low user engagement is.

They would have to vote many more times and it would stick out as a malicious pattern more obviously.

I disagree - my proposal does nothing to impact user experience.

This is a red herring, it’s not what I’m proposing.

Nothing in my proposal makes it harder for bona fide people to vote.

It’s exactly the same to vote. The central premise of my suggestion is totally mainstream and at the heart of most rating systems thaf attempt to make gaming harder.

This is pure opinion. Similar approaches are adopted across the Internet without causing riots.

I ran the numbers for years. The patterns are obvious. Reduce the effectiveness of a 1 vote, and they'll simply cast more. You can predict within a very narrow range when the bombing will stop. It would be even narrower if it wasn't for the numbers the bad actors depend on having infrequent updates. If they updated every hour or so, you could call it on the dot. The end result of your suggestion is is no change, which makes any effort put toward implementing it wasted time. Do daily tracking of a couple of contests and you'll see it yourself. It would be much easier now, since you can create a list with all the contest stories in it and have most of the numbers on a single page absent the vote totals, which you would still have to gather from the 7-day toplist.

You're deluding yourself if you think authors are the only ones manipulating scores. The fans are far more numerous, and to pretend there's not a contingent of them that are borderline rabid in their worship is naive in the extreme. The rest are prone to "innocently" helping with the odd double five or one bomb, which rapidly accumulate when done en masse — not through some grand conspiracy, but simple human nature. That's not speculation, it's experience. When I started leaving a closing note on all my stories telling people to cut that shit out, the "innocent" ones did. My number of swept votes during contests plummeted, and I never saw a score dip from sweeps again. I got lazy about it for a while, and sure enough, the number of swept votes jumped, while the score jumps dipped. Knuckled down, and back to minimal swept votes with big jumps at the end. People underestimate the propensity for maleficence in the readership, and especially in their own.

I've been on three different erotica sites for going on 20 years now. The other two have experimented with every scoring system and mathematical manipulation of same under the sun. There's screeching every time there's anything more than a simple average applied. This is not speculation, it's experience. Even here you can see people saying the sweeps are biased against them. It's math they can't understand, and thus witchcraft.

The one and only thing that could be done here which would improve author experience, widen readers' selection criteria to expose them to more work they'll enjoy, and make trolls work harder would be to eliminate the Red H. That artificial bar creates more day-to-day angst than anything else. Wouldn't do squat for the manipulation of the contests or the topists. Nothing will accomplish that. Anything that makes it harder for bad actors to game the system makes it harder for regular users to use it as intended, ( or is a pointless waste of time ) and the average joe is far less likely to jump through hoops than some obsessed troll. The end result is votes from bad actors having more weight, because fewer votes come from the average joe.

 

[EmilyMiller]

I ran the numbers for years. The patterns are obvious. Reduce the effectiveness of a 1 vote, and they'll simply cast more. You can predict within a very narrow range when the bombing will stop. It would be even narrower if it wasn't for the numbers the bad actors depend on having infrequent updates. If they updated every hour or so, you could call it on the dot. The end result of your suggestion is is no change, which makes any effort put toward implementing it wasted time. Do daily tracking of a couple of contests and you'll see it yourself. It would be much easier now, since you can create a list with all the contest stories in it and have most of the numbers on a single page absent the vote totals, which you would still have to gather from the 7-day toplist.

You're deluding yourself if you think authors are the only ones manipulating scores. The fans are far more numerous, and to pretend there's not a contingent of them that are borderline rabid in their worship is naive in the extreme. The rest are prone to "innocently" helping with the odd double five or one bomb, which rapidly accumulate when done en masse — not through some grand conspiracy, but simple human nature. That's not speculation, it's experience. When I started leaving a closing note on all my stories telling people to cut that shit out, the "innocent" ones did. My number of swept votes during contests plummeted, and I never saw a score dip from sweeps again. I got lazy about it for a while, and sure enough, the number of swept votes jumped, while the score jumps dipped. Knuckled down, and back to minimal swept votes with big jumps at the end. People underestimate the propensity for maleficence in the readership, and especially in their own.

I've been on three different erotica sites for going on 20 years now. The other two have experimented with every scoring system and mathematical manipulation of same under the sun. There's screeching every time there's anything more than a simple average applied. This is not speculation, it's experience. Even here you can see people saying the sweeps are biased against them. It's math they can't understand, and thus witchcraft.

The one and only thing that could be done here which would improve author experience, widen readers' selection criteria to expose them to more work they'll enjoy, and make trolls work harder would be to eliminate the Red H. That artificial bar creates more day-to-day angst than anything else. Wouldn't do squat for the manipulation of the contests or the topists. Nothing will accomplish that. Anything that makes it harder for bad actors to game the system makes it harder for regular users to use it as intended, ( or is a pointless waste of time ) and the average joe is far less likely to jump through hoops than some obsessed troll. The end result is votes from bad actors having more weight, because fewer votes come from the average joe.

I completely disagree. I don’t think your points are valid. But there is only one way to find out

I realize the system works for you, but it doesn’t for many people.

I have my opinion, you have yours.

 

[RejectReality]

I completely disagree. I don’t think your points are valid. But there is only one way to find out

I realize the system works for you, but it doesn’t for many people.

I have my opinion, you have yours.

The system doesn't work for ANYBODY. Neither does the one on SOL or the one on Lush or the one on Amazon or the one on Rotten Tomatoes. They're all plagued with manipulation. That's the nature of an internet poll. It only requires a minimum of effort on Google to find a way around any reasonable control that even a drooling moron can implement.

The one and only way to reduce the impact of malicious voting is to keep plugging away, producing good work, and growing your following. ( And telling them not to engage in scumbaggery consistently )

 

[EmilyMiller]

The system doesn't work for ANYBODY. Neither does the one on SOL or the one on Lush or the one on Amazon or the one on Rotten Tomatoes. They're all plagued with manipulation. That's the nature of an internet poll. It only requires a minimum of effort on Google to find a way around any reasonable control that even a drooling moron can implement.

The one and only way to reduce the impact of malicious voting is to keep plugging away, producing good work, and growing your following. ( And telling them not to engage in scumbaggery consistently )

You can keep saying the same things and continue to fail to convince me, or you can give up on trying to educate the dumb bitch. Your call.

 

[Gamblnluck]

I don't usually check the ratings for my stories, but I happened to glance at the page today and saw something weird. A few weeks back, one of my stories had a rating of something like 4.38, and today the same story has a 4.57 rating on the same number of ratings as it had before. Has this happened to anyone else?

After a few weeks on a well received story, I question why nobody else would rate it. I notice rating fluctuations myself on occasion with little change in numbers. I think that is because some readers go back just to read comments others have left. (That increases the view count as well.) A few of those may change their original rating.
I've had comments that told me that they changed their score up and down after re-reading. Especially after a subsequent chapter was posted.

 

[lovecraft68]

This is pure opinion. Similar approaches are adopted across the Internet without causing riots.

He's been here 19 years (his other ID is 2006) and no one knows the site better.

There's been a lot of ideas about changes over the years, some with more merit than others, but end of the day? Its not about what would or wouldn't work, its about the site not wanting to change. 25+ years in, they're not interested in doing it any differently.

Its the playing field we're part of and this forum is a tiny representation of the writers here and I have a feeling most don't care.

 

[SissyBrandi19]

I don't know, but you raise a fascinating point. As a professional and highly knowledgeable statistical analyst I can assure people that the mean value of a set of data cannot change unless the data itself changes. Now, can people who rate a story go back and change their original rating? I.e. no data added or deleted but one or more data points altered? I don't know if that is possible. But the shift you mention - how many ratings is that based on? Because if you have, IDK, 50 or more ratings that went into the original score you report, that "revised" score is a HUGE shift. But if the scores are based on very few ratings (data points) then it's not such a huge shift. That's why once a mean (which is a type of average, along with median and mode) is based on a sufficient number of scores, it will not shift very much at all. Your bowling average isn't going to change a lot even with a 300 game if it's based on 100 games versus 10 games.