Weird rating goings-on

I can see evidence of sweeps on my stories in E/V and EC in Nov and Dec last year.

Perhaps the way they do it does not cover all stories in a category, or the downvotes on yours did not trigger an elimination.
All the evidence I've seen over the years is that every sweep goes through every category, and goes a long way back in time. I've seen some of my ten years old stories have a scores wobble, occasionally.
 
So you believe, or think you've seen evidence, that good ratings are removed in sweeps, for whatever reason? First I've heard of this idea. I guess I could see it, if the account showed other evidence of clearly unfair votes on ratings: just throw out all votes from that account, essentially tossing out the baby with the bathwater?
A throw-away comment from AHer a long time ago proved it for me. He bragged that he'd gone to one of his stories with an Alt, gave himself a five, but later observed (not seeing the connection himself) that the five disappeared in the next sweep.

I then did an experiment on one of my own stories, gave myself an anonymous five by jumping to the end. Same thing, next sweep through, it was removed. Told me all I needed to know.
 
The objective of a sweep is to eliminate any votes not given in good faith. These could be a five intended to boost a story's position on top lists or in contests. In other words, not all 5s are good, just as some 1s may be genuinely felt.
I sort of get what you're saying, but the question would be what criteria would you use to say a high score was offered in bad faith? It's not hard to see that a '1' score on a toplist story is very probably not genuine. But if someone gives it a '5', without having read it, just to artificially boost it...how would you tell?
 
A throw-away comment from AHer a long time ago proved it for me. He bragged that he'd gone to one of his stories with an Alt, gave himself a five, but later observed (not seeing the connection himself) that the five disappeared in the next sweep.

I then did an experiment on one of my own stories, gave myself an anonymous five by jumping to the end. Same thing, next sweep through, it was removed. Told me all I needed to know.
Is it possible Lit identified the Alt as being from your computer's URL, and thus KNEW it was your own alt? That could explain AHer's experience also.
 
I sort of get what you're saying, but the question would be what criteria would you use to say a high score was offered in bad faith? It's not hard to see that a '1' score on a toplist story is very probably not genuine. But if someone gives it a '5', without having read it, just to artificially boost it...how would you tell?
It has some ability to recognize when the same person is voting more than once on a work, even if that person is logging out to send anonymous votes.
It has some ability to recognize when someone is sending out an unrealistic number of votes in a period of time.
It has some ability to recognize when someone is opening a story just to leave a rating.

It's not perfect, and the mechanisms it uses to spot these kinds of behavior should not be openly speculated on, let lone divulged. Because obviously, if you knew what the exact criteria were you could spoof them easily by slightly adjusting behavior.
 
I sort of get what you're saying, but the question would be what criteria would you use to say a high score was offered in bad faith? It's not hard to see that a '1' score on a toplist story is very probably not genuine. But if someone gives it a '5', without having read it, just to artificially boost it...how would you tell?
I don't know how a sweep works, but I guess looking at the IP address is one factor.
 
It has some ability to recognize when someone is opening a story just to leave a rating.
This, especially, ought to flag bad-faith votes. Leaving a vote within a few seconds of opening a story would have to be strong evidence. I'm sure there are ways around it, but many trolls are probably too lazy to bother.
 
All the evidence I've seen over the years is that every sweep goes through every category, and goes a long way back in time. I've seen some of my ten years old stories have a scores wobble, occasionally.
My understanding is that the sweep looks at the vote table and includes all votes cast since the last sweep. I don't know if that's true, but it's consistent with everything I've read about the sweeps.
 
Is it possible Lit identified the Alt as being from your computer's URL, and thus KNEW it was your own alt? That could explain AHer's experience also.
Other experiments on my own stories suggest a multiplicity of things get looked at with the algorithm, not just one thing. Identity of the source computer is obvious, but I think there's are several other things being looked at. I won't say any more.
 
It has some ability to recognize when the same person is voting more than once on a work, even if that person is logging out to send anonymous votes.
It has some ability to recognize when someone is sending out an unrealistic number of votes in a period of time.
It has some ability to recognize when someone is opening a story just to leave a rating.

It's not perfect, and the mechanisms it uses to spot these kinds of behavior should not be openly speculated on, let lone divulged. Because obviously, if you knew what the exact criteria were you could spoof them easily by slightly adjusting behavior.
Which is why open speculation is frowned upon. If you think you know how it works, best keep it to yourself. I know for sure it covers more than Contest stories, because I don't enter Contests, yet benefit whenever a sweep goes through.
 
So you believe, or think you've seen evidence, that good ratings are removed in sweeps, for whatever reason?

I have screenshots of someone's contest story from a few years ago that lost half its votes overnight, most of them 5s.

The story had 161 votes with a score of 4.86. The next morning it had 80 votes with a score of 4.83.

It's not anyone who posts to AH regularly, and it doesn't mean the author had any involvement in or knowledge of what happened. But sweeps of fraudulent five-star votes definitely happen.
 
Which is why open speculation is frowned upon. If you think you know how it works, best keep it to yourself. I know for sure it covers more than Contest stories, because I don't enter Contests, yet benefit whenever a sweep goes through.
Yeah, it doesn't protect any particular story, it removes ratings that meet one of the criteria for being suspected of being inauthentic.

This has the largest impact on contest entries, because that is where the most inauthentic ratings are made, and especially where the inauthentic ratings are being registered by people who don't know what they are doing. And also because one of the main purposes of fake ratings in contests (artificially inflate or deflate scours for the purposes of driving up or down viewership while the story is on the new releases page) can be successful even if the troll votes get pruned before the end of the contest. If trollbombing a competitor's story deprives them of thousands of views in their first couple of days, it doesn't really matter if those votes are later removed - the engagement lost can never be recovered and the work can't ever get enough genuine ratings to win the contest. Conversely, a bunch of fake 5^ ratings on a contest entry can get it a Red H right away, which will translate to extra readers and real ratings even if the bogus ratings are later pruned.

The trolls who curate the toplists, on the other hand, need to keep their votes in for the long haul. But they obviously know how to evade the sweeps. Whatever they are doing is working, because otherwise they would not be able to maintain plateaus with hundreds of stories hardcapped at the ratings they allow.
 
Yeah, it doesn't protect any particular story, it removes ratings that meet one of the criteria for being suspected of being inauthentic.

This has the largest impact on contest entries, because that is where the most inauthentic ratings are made, and especially where the inauthentic ratings are being registered by people who don't know what they are doing. And also because one of the main purposes of fake ratings in contests (artificially inflate or deflate scours for the purposes of driving up or down viewership while the story is on the new releases page) can be successful even if the troll votes get pruned before the end of the contest. If trollbombing a competitor's story deprives them of thousands of views in their first couple of days, it doesn't really matter if those votes are later removed - the engagement lost can never be recovered and the work can't ever get enough genuine ratings to win the contest. Conversely, a bunch of fake 5^ ratings on a contest entry can get it a Red H right away, which will translate to extra readers and real ratings even if the bogus ratings are later pruned.

The trolls who curate the toplists, on the other hand, need to keep their votes in for the long haul. But they obviously know how to evade the sweeps. Whatever they are doing is working, because otherwise they would not be able to maintain plateaus with hundreds of stories hardcapped at the ratings they allow.
Honestly, this is ass-backwards.

So long as a story still has 25 votes after the sweeps, troll bombs actually serve to insulate a story from further ( and perhaps more sophisticated ) bombing, as well as liked it, didn't love it 4s. It essentially camoflauges them beause the score is below the threat threshold, leaving the story with primarily follower's 5s. 25 votes is barely an ask outside of a few very light readership categories, and those only to fairly new authors with next to no following.

What ends up happening more often than not is that a story carrying a 4.9ish average with sub-50 votes gets bombed just a couple of times to drop into the 4.7s, where it will be ignored by anyone attempting to manipulate the results, because it's well below the leaders, which are now the threat to their favorite story/author. The end of the contest hits, the sweeps roll through, those couple of bombs go away, and the story comes seemingly out of nowhere to take one of the top 3 spots.
 
Honestly, this is ass-backwards.

So long as a story still has 25 votes after the sweeps, troll bombs actually serve to insulate a story from further ( and perhaps more sophisticated ) bombing, as well as liked it, didn't love it 4s. It essentially camoflauges them beause the score is below the threat threshold, leaving the story with primarily follower's 5s. 25 votes is barely an ask outside of a few very light readership categories, and those only to fairly new authors with next to no following.

What ends up happening more often than not is that a story carrying a 4.9ish average with sub-50 votes gets bombed just a couple of times to drop into the 4.7s, where it will be ignored by anyone attempting to manipulate the results, because it's well below the leaders, which are now the threat to their favorite story/author. The end of the contest hits, the sweeps roll through, those couple of bombs go away, and the story comes seemingly out of nowhere to take one of the top 3 spots.
Not far from the truth. My Winter Holiday story opened at 4.9 but by day 2 was at 4.64. Quietly out of sight, it got 10 days of nothing but 5s (if my math is right), unexpectedly climbing over 4.8 on the final day of voting but not by enough to give me hope. Then the brooms swept me up to win third place. Thank you, readers! And thank you, sweeps!
 
Not far from the truth. My Winter Holiday story opened at 4.9 but by day 2 was at 4.64. Quietly out of sight, it got 10 days of nothing but 5s (if my math is right), unexpectedly climbing over 4.8 on the final day of voting but not by enough to give me hope. Then the brooms swept me up to win third place. Thank you, readers! And thank you, sweeps!
I was in a very similar boat. For Christmas I opened at ~4.9, quickly got hammered to a ~4.6, climbed back up to a little over 4.7, then after the sweeps and the contest was over it ended up at 4.86, which was right at the cusp of placing (assuming that was the score being evaluated when they were determining winners, but I have no idea when they did that and what my score was at the time of evaluation). I'm currently in a similar boat with my Valentine's Day story. Soon as it breached 4.8, I got whacked to a 4.61. Though to be fair, it has very few ratings (currently just 17, so not even qualifying yet) and it could have been a couple legit 1s (niche anthro erotica isn't for everyone, neither are tamanduas). The low number of ratings makes me highly susceptible to bombings, so if any of my 1s get swept, it'll definitely be a big bounce back.

Also, I really liked your story. :heart:
 
Wait! I just got 4-bombed! 😜
Why waste such an important and tragic post in a long thread many have gotten tired of reading, or at least taking seriously.

I suggest immediately starting a thread specifically about this awful event and demanding rapid and decisive action from the site to put an end to any vote not a five being allowed on any of the stories here.

People will line up to support you, some will even call you brave as you courageously came forward to expose the terrible underbelly of the readership here who on average votes most stories to scores well over four out of five.
 
Back
Top