Vote Culling Data

[designatedvictim]

My Nude Day story went from 4.51 with 99 votes to 4.68 with 92 votes. I didn't expect it to win, but it seems like someone was trying very hard to make sure.

Wow... I was planning to say that the biggest one-increment drop I've seen was back on 3/29, was three, but seven takes the cake!

Not everyone gets a vote dinged on any one sweep - and it's always possible to lose vote(s) to a sweep and gain legit ones for no net change - but my last net-lost vote was about 45 hours ago.

 

[MelissaBaby]

They were announced yesterday.

Simon's little wheel has been spinning for a long time.

 

[ElectricBlue]

There is often one several days before the end of a contest, then one just before it closes, so a sweep on July 3-4 makes sense. Five Inch Heels link seems to confirm that.

Definitely something in the last few days. My last story .jumped from 4.56 to 4.74, with not much of change in votes.

 

[THBGato]

The downside is when your score jumps up too high so it becomes troll-bait and immediately gets pummelled again.

I had one at 4.86. Perfect. In the top list, but somewhere in the relative obscurity of the fourth page. Then a sweep happened and it went up to 4.88, into the top ten. For about 3 hours. Now it's 4.83...

Swings and roundabouts.

(Not a contest story.)

 

[Five_Inch_Heels]

Yes, becoming a target gets arrows slung.

 

[onehitwanda]

You didn't think it was going to escape the Hangout unscathed, did you?

Unscathed, no. Still functional... well, a girl can dream, can't she?

 

[SimonDoom]

Unscathed, no. Still functional... well, a girl can dream, can't she?

It doesn't hurt. I'm keeping a watchful eye on mine. It's going to be a close call whether we both make it out of here. It doesn't help that the resident doctor is constantly telling us to drink.

 

[onehitwanda]

It doesn't hurt. I'm keeping a watchful eye on mine. It's going to be a close call whether we both make it out of here. It doesn't help that the resident doctor is constantly telling us to drink.

I think she has an ulterior motive.

 

[Gamblnluck]

You would think that if a sweep was done, the numbers wouldn't drop by only 1 to 3. You'd see a significant change. Especially when the previous time frames only went up by 1 or 2.
I'm wondering what happens when I read a story and say give it a 3. I go back later, and re-read it. This second time through, I change my vote to a four. Does that just replace the old 3 or add a new vote of 4 if a few weeks have passed. (I only see the yellow stars.)
At the same time I noticed while I can change my vote up and down, I cannot remove it altogether by clicking to one side.

 

[AALP]

The downside is when your score jumps up too high so it becomes troll-bait and immediately gets pummelled again.

I had one at 4.86. Perfect. In the top list, but somewhere in the relative obscurity of the fourth page. Then a sweep happened and it went up to 4.88, into the top ten. For about 3 hours. Now it's 4.83...

The most frustrating part of the current bombing of the top lists is that it's being done on an automated basis by more than one party, which means it's highly unlikely to stop.

However, it's not just one-bombing that's a problem here, but fake five-star ratings.

A while ago, an author on the story side of Literotica boasted about awarding multiple fake five-star ratings. I'm not going to name them here, but I will share their data as they posted it publicly and how it affects their own stories.




The table above is straightforward, apologies for formatting. P = Pages, PV = Page Views and V = Votes Cast. For Story 6, there are 9 Literotica pages. So, for a legitimate vote to be cast, a reader must read all pages, which for Story 6 comes out as 9 page views per genuine read.

The maximum possible number of genuine votes can be calculated as page views divided by the number of pages. In the case of Story 6, the table shows that the maximum possible number of genuine votes is 588. However, that author's data shows they gained 446 total votes, which in percentage terms is a staggering 75.75% of the maximum possible number of genuine votes (588).

Vote-to-maximum-readership (VtMR) rates above 50% are statistically implausible on Literotica as the vast majority of readers do not vote, with a VtMR percentage of 20% being very strong. We all know there are lots of reasons why readers don't sit down and click each page in turn to finish a story. Even massively popular stories rarely exceed a VtMR percentage of 15%. My friend's story, which was very popular here until it got published and held the #1 position in the category's top list, had a VtMR rate of only 3.3%.

The table above shows all but one story scoring massively above that 15% benchmark, with red warnings on scores above 50% that are so inflated beyond the limits of credibility that they're basically taking the piss.

Not every cheater on the website goes to those amazing lengths, but it's far more widespread than anyone wants to admit. That's why I recommend that no one feels bad about their own honest scores, because some of the people you think you're competing with are not playing fair.

 

[alohadave]

You would think that if a sweep was done, the numbers wouldn't drop by only 1 to 3. You'd see a significant change. Especially when the previous time frames only went up by 1 or 2.
I'm wondering what happens when I read a story and say give it a 3. I go back later, and re-read it. This second time through, I change my vote to a four. Does that just replace the old 3 or add a new vote of 4 if a few weeks have passed. (I only see the yellow stars.)
At the same time I noticed while I can change my vote up and down, I cannot remove it altogether by clicking to one side.

If you change your vote, there are still the same number of votes on the story.

I track my stats weekly, and 1-3 votes removed on a story is normal for a sweep. There have been very rare cases where more than that have been removed from a story (IME, YMMV, TINLA).

 

[TheLobster]

So, for a legitimate vote to be cast, a reader must read all pages, which for Story 6 comes out as 9 page views per genuine read.

Don’t we have some pretty strong evidence that only page 1 hits are counted as views? That’s what the consensus seems to be among AH denizens; but if it’s wrong, then this also messes up any comparison of view numbers between stories of different lenghs.

Unfortunately, I don’t believe we have an official definition of what a view is.

 

[NotWise]

If you change your vote, there are still the same number of votes on the story.

I track my stats weekly, and 1-3 votes removed on a story is normal for a sweep. There have been very rare cases where more than that have been removed from a story (IME, YMMV, TINLA).

In 2021, I had 24 votes removed from a Nude Day story, with 18 of them within a two-hour period before the announcement. It bumped the rating from 4.21 to 4.64. There was a time when I could track changes in contest scores, and I saw similar large changes several times. Things may have changed since then.

 

[Bamagan]

Don’t we have some pretty strong evidence that only page 1 hits are counted as views? That’s what the consensus seems to be among AH denizens; but if it’s wrong, then this also messes up any comparison of view numbers between stories of different lenghs.

Unfortunately, I don’t believe we have an official definition of what a view is.

Every time I've tested it, none of the pages except the first has affected the view count.

 

[filthytrancendence]

Don’t we have some pretty strong evidence that only page 1 hits are counted as views? That’s what the consensus seems to be among AH denizens; but if it’s wrong, then this also messes up any comparison of view numbers between stories of different lenghs.

Unfortunately, I don’t believe we have an official definition of what a view is.

This level of trying to dig into it seems more like reading tea leaves than analysis to me. As you say, there's too many assumptions being made.

And ultimately, while the downsides of the public anonymous voting system can be frustrating, you don't have to think about it very hard to see that any cure would be probably be worse than the disease. For 99.9% of stories, the worst you have to worry about is some disgruntled asshat 1-bombing you, and the sweeps are a reasonable way to address this. And it makes sense that it's mostly done around contest time.

I think we just have to accept that the top lists are always going to be subject to some fuckery, because playing whack-a-mole with all the various techniques that might be employed to fuck with top list scores one way or another is simply not the best use of Manu's time.

 

[NotWise]

The table above is straightforward, apologies for formatting. P = Pages, PV = Page Views and V = Votes Cast. For Story 6, there are 9 Literotica pages. So, for a legitimate vote to be cast, a reader must read all pages, which for Story 6 comes out as 9 page views per genuine read.

I disagree with the assertion that a legitimate vote requires the reader to read to the end. Readers don't have take long to decide they hate a story, and that should be a valid vote. I've also seen readers comment that they loved the opening and voted 5* before they read on. That seems less valid to me, but I don't know a good way to distinguish between those cases.

 

[designatedvictim]

The maximum possible number of genuine votes can be calculated as page views divided by the number of pages. In the case of Story 6, the table shows that the maximum possible number of genuine votes is 588. However, that author's data shows they gained 446 total votes, which in percentage terms is a staggering 75.75% of the maximum possible number of genuine votes (588).

I think part of the problem is that the value labelled in the screenshot, called PV and said to be page views, it actually story views.

The downloaded stats file simply names that column View Count.

That means dividing by the page count doesn't apply, as TheLobster said already.

That affects all of your assertions.