Score Vandalism

[OverconfidentSarcasm]

This is not a good idea. Stories Online does something similar to this, and it's basically a scoring shitshow. It also doesn't actually fix the problem, i.e., if someone wants to tank a story's rating, they can shift to making it a 2 or 3 star rating instead.

In all honesty, I never understood why people have a problem with SOL's rating system. Yeah, it's harder to get a 10/10 score unless ALL votes your story receives are 10s, but, other than that, I don't see any disadvantage.

I mean, as far as I understand it, Literotica periodically sweeps scores to get rid of the one-bombs. And, while I fully agree with that practice being necessary because of the trolls... I don't see how that doesn't turn Lit's scores into just as unrepresentative of reality as that weird formula on SOL does.

But I truly do believe that registering to vote would help. After all, yes, people could just create multiple accounts... but I can not believe that ALL the people who like to manipulate scores on this site right now would still be doing it if they had to create multiple accounts for it. Especially since Lit could VERY easily identify those accounts if they never do anything but log in, instantly jump to the last page to one-vote a story, because waiting a reasonable amount of time to make it look like they read it would be contra productive for the whole exercise, and then log out again.

 

[pink_silk_glove]

Has rating breakdown ever been proposed? Like, not change anything, just let authors see how many 5s and how many 1s they have, instead of an average?

I'd love to see that. It also speaks volumes as to how shoddy the site code is, since the the sweeps prove that the data is there, all they have to do is show it.

 

[TheLobster]

I'd love to see that. It also speaks volumes as to how shoddy the site code is, since the the sweeps prove that the data is there, all they have to do is show it.

I would bet that there are large swathes of code under the hood that haven't been touched for longer that some users have been alive.

And yes, I say this knowing it's a site for adults.

 

[AwkwardlySet]

How do you know this is happening? That people are giving stories ten or 20 1 stars? How could you possibly know that?

I've been publishing stories here for over 7 years, and I've received many 1s and negative comments, but I see no clear evidence of this behavior.

I know it because it happened to me and to a lesser it still is happening. I also know it because I tested it myself. In my frustration, I tried to vote fraudulently on stories and was surprised at how easy it was to do it. I was meticulous in my testing and I even waited to see if sweeping would remove such votes. Guess what, the sweep didn't really do its job because the algorithm isn't all that much. There is a logical way around it, one I am not going to say because I don't want to promote this abuse of rules.
Just because you yourself haven't experienced something in your seven years here, doesn't mean it's not happening, or are you just standing in for EB here?

 

[KeithD]

since the the sweeps prove that the data is there

I'm not sure that the sweeps are scientifically sound. But, as noted, the existence of the sweeps shows the perceived need for them because the voting system is faulty.

 

[AwkwardlySet]

In all honesty, I never understood why people have a problem with SOL's rating system. Yeah, it's harder to get a 10/10 score unless ALL votes your story receives are 10s, but, other than that, I don't see any disadvantage.

I mean, as far as I understand it, Literotica periodically sweeps scores to get rid of the one-bombs. And, while I fully agree with that practice being necessary because of the trolls... I don't see how that doesn't turn Lit's scores into just as unrepresentative of reality as that weird formula on SOL does.

But I truly do believe that registering to vote would help. After all, yes, people could just create multiple accounts... but I can not believe that ALL the people who like to manipulate scores on this site right now would still be doing it if they had to create multiple accounts for it. Especially since Lit could VERY easily identify those accounts if they never do anything but log in, instantly jump to the last page to one-vote a story, because waiting a reasonable amount of time to make it look like they read it would be contra productive for the whole exercise, and then log out again.

Sol's system is better, I agree, it's just that people are used to ridiculously high scores here and they complain about the lower scores they get there. For my own stories there, I think I received only about 10% fewer votes than on Lit, which is remarkable considering that only registered users are allowed to vote. SOL has potential in my opinion, but the interface design is atrocious

 

[pink_silk_glove]

They were always meant to be a tool that shows how much readers liked or disliked a story, they were never meant to be a gauge of story quality.

Yes, and still inaccurate because thumbs-up/thumbs-down is all they would need for that. Instead we have this 1-5 scale where everyone has a different idea of what a 5 or a 4 or a 3 is ... etc. On top of that, we have this stupid Red H set as am immobile bar at 4.5 that fucks up everything.

I completely understand your beef with this and long ago I voiced the same objections you are voicing because they are generally true. But since then I managed to accept that scoring isn't what we want it to be. It's what it is, as accurate or inaccurate as that makes it.

And I have accepted that too, more than anyone else on this forum, but the thread is about improvements to the scoring system and so I've offered my suggestions.

After all the experience I had with being bombed for close to a year now, I'll just disagree with you. Sweeps barely scrape the problem and what's more important, sweeps take Laurel's time, which is why she doesn't employ them often enough, not to mention the fact that sweeps use a pretty crude algorithm, one that many persistent trolls find a way around.

You're not disagreeing with me at all. Personally I've been plagued with bombs on several occasions. I get such low traffic that it's easy to spot when every story in my catalog gets precisely one vote and goes down (usually right after someone starts a pissing match with me in the forum or in the chat). I've also been multi account bombed (like dozens of accounts) a couple of times. With my low traffic, I'd bet 1000 dollars that I have the highest bomb ratio of anyone on lit, certainly anyone in the AH. The sweeps hardly catch all of the bombs (how can they?) and they're extra admin to handle. We agree on that 100%. What I'm saying is that a better scoring system would reduce the bombs and reduce the need for the extra admin of sweeps.

 

[pink_silk_glove]

I'm not sure that the sweeps are scientifically sound. But, as noted, the existence of the sweeps shows the perceived need for them because the voting system is faulty.

Oh, the sweeps are far from perfect, but they have to have them, but that's not the point.

All that I am saying is, in order to conduct sweeps the site has to keep track of individual votes. If they have those individual votes, they could easily code on your profile page a vote breakdown (how many 5s, 4s, 3s etc). They have the data but won't show it. They could code it into the profile page in an afternoon.

 

[alohadave]

It will. Any H story with 1000 votes needs only two 1s to drop by .01 in score. This is just the math arithmetic mean, and is one of the reasons practically no one uses such a flawed scoring metric anymore.

In theory the score will change. I have stories that have gotten dozens of votes without the score changing at all. Something happens to make scores 'sticky' so that they don't move up as readily after a while.

 

[Lifestyle66]

The site Admins run sweeps of the ratings for new authors and for those who continually bitch about their ratings or ask for a sweep.

My story rating numbers haven't decreased for months, because I've come to these forum posts and said, "I don't care, and you can't make me care! Do your worst, and 1-bomb the shit out of my stories!" I've even written stories daring the readers to hate it (ie, Pavlov's Dog - 750 Words, at 2.29 with 669 votes in LW)! So, they've stopped running their sweep algorithm against my stories.

For those of you here bitching about the rating system, the site Admins will throw you another bone and sweep the shit under the rug of one or two of your stories to prop it up, maybe even to give you another Red-H ... because you deserve it! For me, i'll anxiously wait to see if Pavlov drops to 2.28 or lower.

 

[MrPixel]

The evidence comes from watching the 1-bombing patterns.

When I had a contentious debate with two authors here in the AH a few years ago, each of my Red-H stories soon (within an hour) received 1s until several dropped below 4.5. Then when another 5 was added to raise one of them to 4.5 Red-H again, within 30 minutes another 1 would push it back down.

My story list was being watched for positive changes, and attacked to bring some of those down, over, and over, for about two days.

I can't believe that was random chance. And it happens, even now, with every new story I post. Watching the ratings progress over the first hours shows patterns of 5s, followed by 1s, back & forth until the attacker grows staisfied and moves on to their next target.

This echoes what I'm seeing, as well. Like this morning. There are folks here on AH who are doing this. Very unseemly, very childish. You heard me. There's the corner. Go stand in it until you get your mind right.

 

[OverconfidentSarcasm]

Oh, the sweeps are far from perfect, but they have to have them, but that's not the point.

All that I am saying is, in order to conduct sweeps the site has to keep track of individual votes. If they have those individual votes, they could easily code on your profile page a vote breakdown (how many 5s, 4s, 3s etc). They have the data but won't show it. They could code it into the profile page in an afternoon.

The big question for me is: How MUCH data do they actually have.

Do they just go in and delete all one-votes, or do they have a database column where they save additional data to identify which one-votes have been made after (seemingly) reading the story and which one-votes have been made straight away to just bomb the score?

 

[pink_silk_glove]

The big question for me is: How MUCH data do they actually have.

Do they just go in and delete all one-votes, or do they have a database column where they save additional data to identify which one-votes have been made after (seemingly) reading the story and which one-votes have been made straight away to just bomb the score?

I have my theories but I won't say them out loud, as they could tip off the bombers as to how to beat the sweep.

 

[VallesMarineris]

I've been 1-bombed too and it's frustrating. But there is a straightforward solution that would not change the basic system and could be implemented with minimal programming:

1) Change the rating feature so that instead of picking from a row of 5 stars, the reader moves a slider from 1 to 5.

2) Under the hood, the rating is rounded to one decimal place, thus 1.0, 1.1, 1.2 . . . 4.8, 4.9, 5.0.

3) Instead of calculating an average, collect the number of times each rating value is picked. Basically build a histogram of the ratings.

4) Report the story's rating as the median of the scores, not the average. Note that this is the same method used for reporting real estate prices. It's why you read about the median price of a home, not the average. The real estate people don't want a single sale of a two million dollar home to skew their customers' perceptions if only 10 houses were sold that month. If we use the median, then 1-bombs have only a small effect: all they do is lower the reported score down one step in the ordered list of scores.

Bonus: having a histogram instead of a single number tells everyone, including Laurel and Manu, a lot more about the stories. For example, a highly polarizing story will have two peaks in the histogram, a low one and a high one.

 

[OverconfidentSarcasm]

For those of you here bitching about the rating system, the site Admins will throw you another bone and sweep the shit under the rug of one or two of your stories to prop it up, maybe even to give you another Red-H ... because you deserve it! For me, i'll anxiously wait to see if Pavlov drops to 2.28 or lower.

I seriously can not imagine Laurel or Manu actually taking the time to read the forum, make a list of all the authors who complain about bad ratings, and then go through the trouble of mending their scores specifically.

If you ask me, they have a cronjob running that does it automatically for all stories on the site, but (hopefully) using some kind of indicator to determine automatically which stories get swept clean and which stories are justified in having a bad score.

Of course, it's very well possible that they just delete one-votes in general, to make the story catalog seem more attractive than it actually is...

 

[Lifestyle66]

... With my low traffic, I'd bet 1000 dollars that I have the highest bomb ratio of anyone on lit, certainly anyone in the AH. ...

Careful with that bet.

You're up against Pavlov's Dog, which may have 12k views, but 669 votes for a 2.29 average rate!

 

[StillStunned]

Just drop the ratings altogether. Instead, show how many readers have favourited the story, perhaps relative to the number of views.

 

[Lifestyle66]

I seriously can not imagine Laurel or Manu actually taking the time to read the forum, make a list of all the authors who complain about bad ratings, and then go through the trouble of mending their scores specifically.

If you ask me, they have a cronjob running that does it automatically for all stories on the site, but (hopefully) using some kind of indicator to determine automatically which stories get swept clean and which stories are justified in having a bad score.

Of course, it's very well possible that they just delete one-votes in general, to make the story catalog seem more attractive than it actually is...

My stories used to change with a sudden drop in the number of votes overnight the first Tuesday-Wednesday after they came off the New list.

That stopped as I said quite a few months ago. I haven't seen any sudden drop in the vote numbers. But it might be their algorithm doesn't run the sweep against stories with high number of votes, and mine lately have exceeded 100 to 200 votes within the first week. (My latest story posted 06/03/2024 has 940 votes.)

 

[pink_silk_glove]

Careful with that bet.

You're up against Pavlov's Dog, which may have 12k views, but 669 votes for a 2.29 average rate!

My first story was at 4.6 with about 20ish votes after about a year. Then someone harassed me in chat on a dozen accounts and I reported them all and got over a dozen accounts banned from chat. A week later someone pointed out to me that my story (it was the only one on my account at the time) was scored 1.8. I checked it out, it was 1.8 with over 50 votes. Someone went onto over 20 accounts to 1-bomb it. A month or so later the sweeps brought it back up to 2.4, nowhere near 4.6, so the sweeps are far from perfect.

However, you cite one story bombed horribly low (in LW no less, go figure) with 669 votes. My entire catalog (9 stories) has about 300 votes, so my bomb ratio is higher than yours.

 

[Lifestyle66]

My first story was at 4.6 with about 20ish votes after about a year. Then someone harassed me in chat on a dozen accounts and I reported them all and got over a dozen accounts banned from chat. A week later someone pointed out to me that my story (it was the only one on my account at the time) was scored 1.8. I checked it out, it was 1.8 with over 50 votes. Someone went onto over 20 accounts to 1-bomb it. A month or so later the sweeps brought it back up to 2.4, nowhere near 4.6, so the sweeps are far from perfect.

However, you cite one story bombed horribly low (in LW no less, go figure) with 669 votes. My entire catalog (9 stories) has about 300 votes, so my bomb ratio is higher than yours.

When you're talking about ratios, that 669 votes averaging to 2.29 shows far more 1s. And that's just one of my shitty stories! More than 10% of my story are rated 3.0 or lower, with more rated just over 3.

But 38 of my 56 stories are in Loving Wives, the bomb magnet!

 

[TarnishedPenny]

The big question for me is: How MUCH data do they actually have.

Do they just go in and delete all one-votes, or do they have a database column where they save additional data to identify which one-votes have been made after (seemingly) reading the story and which one-votes have been made straight away to just bomb the score?

No offence, but Laurel has requested that we not speculate on how the sweeps work, lest the give would-be trolls ideas.

 

[pink_silk_glove]

When you're talking about ratios, that 669 votes averaging to 2.29 shows far more 1s. And that's just one of my shitty stories! More than 10% of my story are rated 3.0 or lower, with more rated just over 3.

But 38 of my 56 stories are in Loving Wives, the bomb magnet!

Bragger. : P

 

[John_Vandermeer]

Just drop the ratings altogether. Instead, show how many readers have favourited the story, perhaps relative to the number of views.

I would do almost the same thing, but have a thumbs up, relative to the number of views.

Alternatively simplify scoring to 1, 2, 3, and only report percentile rank. Not score, not red H, not anything else.

Or report the full histogram of 1, 2, 3, and nothing else. But then many would probably complain that is too complicated.

Any of those would be better than current system.

 

[SimonDoom]

The evidence comes from watching the 1-bombing patterns.

When I had a contentious debate with two authors here in the AH a few years ago, each of my Red-H stories soon (within an hour) received 1s until several dropped below 4.5. Then when another 5 was added to raise one of them to 4.5 Red-H again, within 30 minutes another 1 would push it back down.

My story list was being watched for positive changes, and attacked to bring some of those down, over, and over, for about two days.

I can't believe that was random chance. And it happens, even now, with every new story I post. Watching the ratings progress over the first hours shows patterns of 5s, followed by 1s, back & forth until the attacker grows staisfied and moves on to their next target.

OK, but if this is true, this is proof AGAINST the OP's main point. The problem isn't unregistered voters and anonymous commenters. The problem is the people participating in this forum, who if they are authors MUST be registered members and voters.

I think it's possible that I've become a victim of this, because last fall when it was announced that I'd won a contest in this forum my score immediately plummeted. Was there retaliation? Maybe. But on the other hand my vote total is now about 4 times what it was then and the score remains around where it was after the downvoting, so that tells me this score probably is more representative of what people think of my story than the score I had when I won the contest (that score lasted for about one day).

More importantly, what's the possible solution to that? How would one filter out such votes? Sweeps might do it, but I don't know, or want to speculate, about how sweeps work. What's clear is that eliminating unregistered votes would NOT take care of this problem.

 

[OverconfidentSarcasm]

No offence, but Laurel has requested that we not speculate on how the sweeps work, lest the give would-be trolls ideas.

Oh, so security by obscurity. Not like there has ever been an instance where that didn't work out perfectly for everyone.