Implemented Fairer sorting of lists

[SealJohnson]

Currently, the trending lists, tag rankings and series finder—indeed, pretty much everything on this site—are by default sorted by total views.

Total views is a pretty meaningless stat, and using it as the primary means to sort series is, to be frank, screwing over a lot of series posted here.

Why is it meaningless, you may ask? Because the total number of views depends on the number of chapters posted. And a chapter can be any length.

A series with lots of tiny chapters is going to have many times more total views than one with fewer, larger chapters, even if they have the exact same number of readers and the same number of words written. This puts authors in a difficult position. Either they have to game the system by dividing their chapters into tiny snippets posted separately, or their series are less likely to get discovered, and don't get read.

A somewhat related problem: the current trending list creates a massive snowball effect, where the only series being viewed are those that already have the most views. This kind of thing is a problem everywhere, but the trending lists on other sites such as Royal Road are much more successful at getting readers to view new stories. Here, it's mostly the same titles, week after week.

Many of the stats shown next to series in the various lists are equally unhelpful.

• Number of chapters: Not a valid unit of measurement. A chapter can be any length.
• Chapters per week: Same problem.
• Favourites: Depends on the number of chapters.
• Total views: See above.

Here are some stats that are actually useful for sorting / showing in summaries:

• Average views: Total views / number of chapters.
• Number of words: Unlike number of chapters, this accurately represents how much has been written.
• Words per week: Much more meaningful than chapters per week.
• Number of pages: Many readers don't know how long a 'page' is, but it's certainly more meaningful than number of chapters.
• Rating: A bit iffy sorting by ratings, because they're so easily abused. Only useful if you take into account the total number of ratings as well as the overall score (e.g. something with a single 5-star rating should not beat something with a 4.9 average and 50 total ratings).
• Number of readers: This would probably be my pick for default sorting stat.

 

[Jemini]

those who write novels of 1k or 2k words per chapter do so for a different audience than those who write 3k+ chapters. It's almost laughable to assume that "trending" titles need to be fair, because the audience are who decide what ends up there. yet Jemini repeatedly accused Xian of "gaming the system" because he writes shorter chapters.

I would like to point out, every time in this thread it has been Xian taking it personally and bringing up their own story. I have kept the terms general, with the exception of calling out Reincarnated Vampire as what I consider a positive example.

Xian, of course, comes in because his position would be threatened by any kind of change in the system from its current model. This is the kind of reason why I am not really convinced by how he keeps saying he doesn't really care about the views. I did compare his series directly a few posts back, but that was after he repeatedly took this case personally and brought himself into it several times in the conversation.

It seems to me as though you are attempting to interpret me deliberately in a negative light on this, and I don't appreciate it. I can understand where Xian would naturally feel a little threatened and make self-interested arguments, and since I also post longer chapters I have to cop to being a little self-interested myself. As such, I do not blame him for taking his stance even if I think it's wrong, and I meet him as best I can to present sound and reasonable arguments. However, what you just said was uncalled for. I called you out for being rude before. I do not think you even realize how extremely rude you are. In my case, I bring up sensitive issues that make it hard to avoid hurt feelings, but I try to make some efforts at least to keep up some decorum. What you just did though crossed the line.

 

[lnv]

I never understood why people would waste their time going through a genre they hate just to rate it negatively -.- people like that shouldn't be permitted to rate anything or leave reviews, because they have no intentions of doing so objectively on a series merits or failings, but 100% personal bias.

as for Jemini's wording, its not so much harsh as it is self-entitled. My personal experience with Jem is they read into your response what they wish to see, infer meanings not present in your posts and argue their entitlement down to the letter.

there are issues with people who post for the trending screen time alone, which is an issue in itself and rightly so, however, the novels who choose to write at this length, do so because they know their target audience.

Of course, there is a target audience for short chapters. I know plenty of novels which work best that way like the quick transmigration novels and etc. A chapter should be written based on how "complete" it is as a chapter. From personal experience, my chapters used to be 1500-2000 in volume 1 because that was what it took to get a complete chapter, come volume 2 the chapter size went up to 2000-3000 and in volume 3 that went to 3000-4000 words. Why? Because that is how long it took to relay completeness in a chapter. So it is up to each author to decide what kind of style they think suites their story be it short chapters or long chapters.

those who write novels of 1k or 2k words per chapter do so for a different audience than those who write 3k+ chapters. It's almost laughable to assume that "trending" titles need to be fair, because the audience are who decide what ends up there. yet Jemini repeatedly accused Xian of "gaming the system" because he writes shorter chapters. that is an argument based on personal convenience. additionally claiming that a story that has received a higher rating is more deserving than one with a lower rating. scores are subjective and that series of 4.8 rating may have only 10 ratings, while the 4 rating may be the result of 50+, in this scenario which rating is more deserving? perhaps when a rating of 5 from 1-3 ratings comes along, does it deserve to rank higher? what if there is a single 0.1 difference? how many people even rate a series compared to those who read it?

if there is really such an issue with people "deserving" more screen time based on trends, just artificially induce a split, if a series has an average word count below 2.5k it goes in List A and if it has a higher average it goes to List B. though this defeats the simple point of being impartial to the series being displayed, on anything besides their 'popularity'.

Jemini's complaints are ultimately biased, as is almost anyone else's on the topic, one way or another. they feel the system is robbing them of much 'deserved' screen time. On the sole basis that they chose to write longer chapters.

I agree on the rating system aspect, we all know the trending system is flawed, so why would anyone assume the rating system is somehow any better? I have not gotten around to grasping SH's ratings yet, but on NU, I tend to see 4+ as decent novels and below that as below average. Of course there will always be things that appeal to niche audience but that is how I see things.

The trending system is slated to be fixed but none of the suggestions so far come close to actually making it any fairer than its current state. Ultimately, different people want the system to give a skewed result in their favour. besides, those series who hold steady in the trends are the only series worth making comparisons to, yet people are determined to argue that these shorter series are less deserving of their ranks, going so far as to claim their only redeeming merit is their release frequency, despite not sharing any similarities to their own work.

Someone can only claim this as a factual statement when the story they are complaining about shares enough similarities to their own that the length of chapters or frequency of their releases become the part of the sole deciding difference to how their traffic fares. rather than treating the faster release of shorter chapters as a multiplicative factor based on word count, perhaps they should first compare the average views per chapter that story is receiving. A comparison that is impossible because such a situation would instead reveal a high likelihood of plagiarism to have occurred.

At first, I considered the views * word count to be more fairer, but when you mentioned word manipulation like CN novels, I don't want that either. So I think a system that is based on unique views/week would probably be as fair as it gets. That is where I'd like to see this discussion go, towards ironing out what can be considered as fair as possible. And not in the direction it is going now where we are talking about which is making it personal.

The unique view/week would allow authors to publish both short chapter and long chapters regardless of how much word count or any other manipulation.

will I click on a trending series? occasionally. after reading the synopsis, how likely am I to continue to the 1st chapter? only as likely as the synopsis is to my taste. and on the account I do start reading the series, the instant I decide it isn't worth continuing, you can expect to see a drop off in views for the following chapters compared to the earliest. now how much of an impact will people claim I have made when the gap between total views and total/average chapter views remains?

That kind of depends? Of course there will be fall off from 1st chapter, but a lot of people would continue reading something that doesn't fall below their bottom line, all depends on how picky they are.

No industry is "fair", the creative arts including writing, even less so. Writers don't even get a say in how well received their work is, lest all writers be international award winners simply for putting pen to paper, fingers to keyboard.

Of course, no industry is fair. But SH isn't an industry. SH's goal from my understanding is to create and outlet for original authors. And in that way, it is important to be as fair as possible. Otherwise, intentionally or unintentionally you end up influencing writing. We have already seen what QI has done to CN with their encouraging word counts.

I want to see a system that would treat both those writing short chapters and long chapters equally. I don't want authors to even think about what would get me more views or less views and just go what makes sense for their story. It is the job of the system to make it as far as (humanly) possible.

 

[Jemini]

1) You get featured on NU more often
2) You get better placement on trending
3) You show up more often on the front page listing

You're missing one, and this one is the most egregious of the list.

4) each additional chapter posted the same day allows you to get additional views without additional unique viewers.

This is the advantage I think is tantamount to just straight up cheating. I think your suggestion of having the system count only one view per unique profile per day would help to mitigate this issue. You have some very good suggestions here. I do not thing the #3 item alone would be enough to upset the situation too much.

Actually, simply implementing the system of 1 view count per 1 unique viewer per 1 day would go a long way toward evening out all the unfair advantages. It wouldn't solve all the issues, but it would solve a lot of them.

 

[XianPiete]

Well, I'm sure people would still accuse you, but it would also be with some respect. It would take a ridiculous amount more effort to post multiple 4K word chapters a day. Thing is, if you increased your chapter size to 4K words, it would place you on a level playing field with the longest chaptered stories on the site. If you are still able to make more than one chapter a day on that standard, that's just you putting in the dedicated effort, and there's no arguing with that.

Again, this whole thing of taking my words to mean "longer = better" is a very superficial and incorrect read of the message. The message is shorter = unfair advantage. An unfair advantage that can, as demonstrated by the example of your story and Reincarnated Vampire, make up for close to a full star ratings' difference.

I am currently writing 4,350 words per day give or take. I edit my chapters down to get to that. My raws before I edit are usually closer 2500 words per chapter. I can easily write 8,500 words each day in my current allotted time. I work a full-time job and do several other hobbies if I had the urge to write more I could up that 12,000 words per day easily. The things that you brag about to me are meaningless.

What I had an issue with is using the star system which when you have been at the top of the trending list for a while gets abused by people who give you 1, 2 and 3 stars as a penalty for being a novel above their own. The other issue I have with you is that you keep saying things like Longer = better quality and people who write more than you are simply trying to rig things in their favor. I am not some kind of malicious genius who grabbed a chance to diabolically place my novel at the top of the list and I resent you portraying me as such over and over again.

Now as for a "fair" list as you stated was your true goal, it needs to be random. Completely random. Allow every single story a chance to show up on the front page. Any other way of sorting will be biased by the writing style of the author, by external factors like time of day it was posted and by the popularity of the author.

 

[lnv]

You're missing one, and this one is the most egregious of the list.

4) each additional chapter posted the same day allows you to get additional views without additional unique viewers.

This is the advantage I think is tantamount to just straight up cheating. I think your suggestion of having the system count only one view per unique profile per day would help to mitigate this issue. You have some very good suggestions here.

That would be same as #2. Trending is based on popularity which is based on views. Either way, the unique views/week would address it.

I do not thing the #3 item alone would be enough to upset the situation too much.

I do not think it is that large either, I simply listed all the advantages. The listing was in order of importance. Yes, NU featuring is worth more than trending (for now, once SH gets bigger it won't be an issue but I'm willing to bet there is more to be gained from NU featuring today)

Actually, simply implementing the system of 1 view count per 1 unique viewer per 1 day would go a long way toward evening out all the unfair advantages. It wouldn't solve all the issues, but it would solve a lot of them.

I thought about that at first, but some authors release 8k-13k chapters 2-3 times a week

What I had an issue with is using the star system which when you have been at the top of the trending list for a while gets abused by people who give you 1, 2 and 3 stars as a penalty for being a novel above their own.

The issue of people rating things based on their prejudices is also important to address fairness. What do you think about the anidb Bayesian rating idea where it would weight the ratings based on how they vote. So if a person is using an account just to down vote people, their weight would be less. Of course this will also work the other way around of people giving 5 stars to everything.

Now as for a "fair" list as you stated was your true goal, it needs to be random. Completely random. Allow every single story a chance to show up on the front page. Any other way of sorting will be biased by the writing style of the author, by external factors like time of day it was posted and by the popularity of the author.

A featured random novel might be an interesting idea. Have a thing where if some stories meet some minimum requirements, they can get randomly featured for the day. But that wouldn't exactly be a trending list...

 

[XianPiete]

A featured random novel might be an interesting idea. Have a thing where if some stories meet some minimum requirements, they can get randomly featured for the day. But that wouldn't exactly be a trending list...

The issue seems to be that people are afraid their works aren't being seen, they are looking for ways to "game the system" to get their "much better" novel above people who write a lot more than they do. I say do away with the trending list and put up a random Novel list. That is absolutely fair and when people aren't getting readers they can blame their own writing instead of another author's work or style.

 

[Jemini]

I am currently writing 4,350 words per day give or take. I edit my chapters down to get to that. My raws before I edit are usually closer 2500 words per chapter. I can easily write 8,500 words each day in my current allotted time. I work a full-time job and do several other hobbies if I had the urge to write more I could up that 12,000 words per day easily. The things that you brag about to me are meaningless.

What I had an issue with is using the star system which when you have been at the top of the trending list for a while gets abused by people who give you 1, 2 and 3 stars as a penalty for being a novel above their own. The other issue I have with you is that you keep saying things like Longer = better quality and people who write more than you are simply trying to rig things in their favor. I am not some kind of malicious genius who grabbed a chance to diabolically place my novel at the top of the list and I resent you portraying me as such over and over again.

Now as for a "fair" list as you stated was your true goal, it needs to be random. Completely random. Allow every single story a chance to show up on the front page. Any other way of sorting will be biased by the writing style of the author, by external factors like time of day it was posted and by the popularity of the author.

On the star system, fair point. I'm trying to get to a fairer system, and discussion like this where someone poses ideas and other people point out the flaws in those ideas. In this manner, Inv has come up with some very good points, and we together have arrived at the far more reasonable solution of counting only 1 view per viewer per day, ingoring multiple chapters viewed by the same viewer.

Once again, there is a difference between having short chapters, posting multiple chapters per day, and the point where the two intersect. When you do both, it is hard to see it any other way than gaming the system. You claim to not care about the views, but if that's the case, then why do you post multiple chapters per day AND write short chapters? I can tell you are getting frustrated, and it's hard to interpret this as anything other than you feeling your position is threatened.

Also, I have addressed the issue of the Longer = Better thing several times already. That's not what I'm saying, and I'm getting a little tired of clarifying it over and over again, and this is another thing that's starting to make me see your arguments as a little disingenuous. I would really like for you to prove me wrong on that notion so we can maybe move on to some more productive conversation.

It's not Longer = Better. It's Shorter = Unfair advantage. An unfair advantage that can have a lower quality 1K word/chapter story beat a higher quality 4K word/chapter story. It's not a level playing field. I'm saying the playing field should be leveled.

Do I even need to say that I'm not saying your story is low quality? It seems like I do since you are taking this very personally.

 

[XianPiete]

You claim to not care about the views, but if that's the case, then why do you post multiple chapters per day AND write short chapters?

I seriously don't care about having first place, I do enjoy seeing my novel on the front page as most authors do. As for your question, I type about 50 words per minute give or take depending on if I have my thoughts in order. I can bang out about 2000-2400 words in an hour. I usually only spend about 45 mins typing then I edit, then I type some more, then I edit. When I post my chapter it's when I feel like it is at the perfect place to end it. Then I start planning my next chapter or I get involved in one of my other hobbies, or maybe I read? It depends. Anyway, I usually feel like coming back to write later which is why sometimes I post two chapters in a row (usually 1-2 hours apart). Or I do like today and get lost on other things and forget to write the next chapter.

If you look at the times of day I post you'll find it is completely random depending on what I am doing that day. My work days I write in the morning here before I work, and again at night when I get home from work. It's not malicious or planned to get views, it's just how I have time and how I want to use my time.

 

[Jemini]

I seriously don't care about having first place, I do enjoy seeing my novel on the front page as most authors do. As for your question, I type about 50 words per minute give or take depending on if I have my thoughts in order. I can bang out about 2000-2400 words in an hour. I usually only spend about 45 mins typing then I edit, then I type some more, then I edit. When I post my chapter it's when I feel like it is at the perfect place to end it. Then I start planning my next chapter or I get involved in one of my other hobbies, or maybe I read? It depends. Anyway, I usually feel like coming back to write later which is why sometimes I post two chapters in a row (usually 1-2 hours apart). Or I do like today and get lost on other things and forget to write the next chapter.

If you look at the times of day I post you'll find it is completely random depending on what I am doing that day. My work days I write in the morning here before I work, and again at night when I get home from work. It's not malicious or planned to get views, it's just how I have time and how I want to use my time.

Well, be that the case or not, (I will choose to believe you,) whether or not that is intentional it gives you an unfair advantage. Again, I am not attacking you personally, I do try to leave your name and story title out of it when I can. However, your story does happen to illustrate the current flaw in the system.

Try taking a chapter of your story that was posted approximately a day ago and look at the view count of that one chapter. Then, look over to the view count on a chapter posted at around the same time on a 4K word per chapter story. If you were to use Reincarnated Vampire for this example (which is something of a power house, but also proves the point,) you will see that the view count for those two chapters posted at around the same exact time as each other is about 3X that of yours.

Again, not desparaging your story, but since it has so toroughly been brought up now it really does serve to illustrate the point. I would say the evidence shows that Reincarnated Vampire is the one actually getting the larger amount of love from the fans. Despite this, Ninetailed-Furball had grown a little distressed and posted a poll to ask if they should shorten the chapters because they also noticed the system was disfavoring them despite the fact that their story was obviously very high quality and very well loved.

If we were to bring my story into this on the same standard, I believe my view count per chapter in a 24 hour time frame is almost identical to yours. Going by this standard, it would mean that our writing is probably at or near the same quality as one another.

As you can see, there actually is a pretty large advantage to be had in posting multiple shorter chapters per day in terms of the trending list as it is currently designed, and that is not completely fair. The counting of 1 view per unique viewer per day would go such a long way toward leveling this out that I honestly doubt anything else would need to be done. You can continue posting as you like, and if that's what the viewers like then you will still have your story doing well. It will do nothing to change your 24 hour view count on a single chapter, which I think is a far better indicator of the story being liked than total view count given the different factors here.

 

[Myriadfold]

Of course, there is a target audience for short chapters. I know plenty of novels which work best that way like the quick transmigration novels and etc. A chapter should be written based on how "complete" it is as a chapter. From personal experience, my chapters used to be 1500-2000 in volume 1 because that was what it took to get a complete chapter, come volume 2 the chapter size went up to 2000-3000 and in volume 3 that went to 3000-4000 words. Why? Because that is how long it took to relay completeness in a chapter. So it is up to each author to decide what kind of style they think suites their story be it short chapters or long chapters.




I agree on the rating system aspect, we all know the trending system is flawed, so why would anyone assume the rating system is somehow any better? I have not gotten around to grasping SH's ratings yet, but on NU, I tend to see 4+ as decent novels and below that as below average. Of course there will always be things that appeal to niche audience but that is how I see things.






At first, I considered the views * word count to be more fairer, but when you mentioned word manipulation like CN novels, I don't want that either. So I think a system that is based on unique views/week would probably be as fair as it gets. That is where I'd like to see this discussion go, towards ironing out what can be considered as fair as possible. And not in the direction it is going now where we are talking about which is making it personal.

The unique view/week would allow authors to publish both short chapter and long chapters regardless of how much word count or any other manipulation.




That kind of depends? Of course there will be fall off from 1st chapter, but a lot of people would continue reading something that doesn't fall below their bottom line, all depends on how picky they are.




Of course, no industry is fair. But SH isn't an industry. SH's goal from my understanding is to create and outlet for original authors. And in that way, it is important to be as fair as possible. Otherwise, intentionally or unintentionally you end up influencing writing. We have already seen what QI has done to CN with their encouraging word counts.

I want to see a system that would treat both those writing short chapters and long chapters equally. I don't want authors to even think about what would get me more views or less views and just go what makes sense for their story. It is the job of the system to make it as far as (humanly) possible.

the unique views would be ideal, based on chapters and not the series as a whole, but the limitations are in how to achieve a proper representations. IPv4 leaves a lot to be desired and while IPv6 is more accurate, its not perfect. the IP address changes and so do the people using it. it can't be based via users because there are far more readers who see no need to make an account and so we slowly run out of factors to determine the unique views as flaws appear for each method that harm the accuracy and thus fairness.

my only issue with unique views is the "re-readability" factor novels, published or web-alike, possess. when a single persons view counts just once no matter how many times they opted to re-read the series, either because later chapters make them rethink what they read in the early chapters, or because they liked a story enough to refresh their memory, though you would hope this isn't happening due to a necessity for confirming the story-verse's facts and continuity. this isn't a key factor to consider in trending as a whole, but it is an indicator worth taking note of.

My own upload being what it was, has 850 views total, only 179 of these are chapter views, bringing my average to 90 views per chapter. I would honestly find it useful to know how many of these are unique in addition to the totals, because it lets me know how many people my work has really reached and how much attention they have paid to it. but this won't be accurate if the same person without an account reads my story multiple times from different locations bringing the validity of the results into question.

my phone has a different IP than all my other internet capable devices. just viewing it again from one of them will inflate the results and doesn't really indicate that the series is "trending" more so than before. same if my device gets reassigned an IP address between viewings.

as for the fall off, for example we can take stats from the top trending series as viewed on the homepage. latest view count will be taken from the most recent chapters averaged to account for deviation. as can be seen below, the longer the series, the less people who actual read it through, this should be a key focus in determining the genuine popularity of a series. this displays a DISADVANTAGE to posting MORE CHAPTERS. As compared to the assumed "advantage" it has gained by being posted more often (same number of people will view the front page per day regardless of how many chapters are present on it). because ultimately there is a lesser completion rate. Yet someone keeps trying to claim that popularity = quality in terms of views, despite the fact that a high quality piece can fail to find readers and a low quality piece can find a fanbase. arguing quality on view count and word count is redundant because it has nothing to do with actual quality.

rank	Chapter View Weight(%)	Chapters(#)	Avg Chapter Views(#)	View Ratio (beginning : recent)	Avg Word Count	Release Frequency (Chapters/week)	Rating	Total Chapter Views
1	75.7%	129	1303	19:1	1269	24	4.1	168126
2	88.9%	340	1845	15.5:1	2331	4	4.3	627539
3	61.2%	21	876	2.6:1	1218	6	4.2	18400
4	63.5%	27	1707	12:1	3932	6	3.7	46095

 

[Jemini]

my only issue with unique views is the "re-readability" factor novels, published or web-alike, possess. when a single persons view counts just once no matter how many times they opted to re-read the series, either because later chapters make them rethink what they read in the early chapters, or because they liked a story enough to refresh their memory, though you would hope this isn't happening due to a necessity for confirming the story-verse's facts and continuity. this isn't a key factor to consider in trending as a whole, but it is an indicator worth taking note of.

That re-read issue is actually pretty easy. Make it on a per-day basis (as I said before.) If it is not on a per-day basis, then that means your rating is entirely based on attracting new viewers and makes no difference if you have a huge loyal following. That would be useless. The per-day basis per unique viewer really just seems to be the most fair system I can think of.

The reason here, in addition to smoothing out the issue with multi-posts per day, it will also even things out when a writer has a HUGE series that already has hundreds of chapters, giving more of a chance to lower chapter count writers who have a lot of viewers liking their smaller number of chapters.

People with a large number of chapters will still have a slight advantage if their series is so long it takes new readers days in order to binge their series, but if they are binging their series then that means they like it, so I think that's quite fair. It would also mean that they are attracting new viewers, witch is another fair marker of being a good series. More to the point, the number of days it would take to binge a large series is more dependent on word count than number of chapters, so this is yet another area in which unique views per day would normalize the large Vs. small chapter discrepancy.

 

[Phantomheart]

I find the current trending and popularity system unfavorable as well, but I don't have that much of a defined opinion on it. Well, rather than say defined, it's more based off the fact that I don't look that much into statistics, but based from face knowledge (seeing the same stories on the site trending everyday with their only common trait being daily/ extremely frequent updates) I see that it's unfair. I write medium length chapters averaging out at 2500 for my bi-weekly novel UD:SG , which is my side of the story.

the unique views would be ideal, based on chapters and not the series as a whole, but the limitations are in how to achieve a proper representations. IPv4 leaves a lot to be desired and while IPv6 is more accurate, its not perfect. the IP address changes and so do the people using it. it can't be based via users because there are far more readers who see no need to make an account and so we slowly run out of factors to determine the unique views as flaws appear for each method that harm the accuracy and thus fairness.

my only issue with unique views is the "re-readability" factor novels, published or web-alike, possess. when a single persons view counts just once no matter how many times they opted to re-read the series, either because later chapters make them rethink what they read in the early chapters, or because they liked a story enough to refresh their memory, though you would hope this isn't happening due to a necessity for confirming the story-verse's facts and continuity. this isn't a key factor to consider in trending as a whole, but it is an indicator worth taking note of.

My own upload being what it was, has 850 views total, only 179 of these are chapter views, bringing my average to 90 views per chapter. I would honestly find it useful to know how many of these are unique in addition to the totals, because it lets me know how many people my work has really reached and how much attention they have paid to it. but this won't be accurate if the same person without an account reads my story multiple times from different locations bringing the validity of the results into question.

my phone has a different IP than all my other internet capable devices. just viewing it again from one of them will inflate the results and doesn't really indicate that the series is "trending" more so than before. same if my device gets reassigned an IP address between viewings.

as for the fall off, for example we can take stats from the top trending series as viewed on the homepage. latest view count will be taken from the most recent chapters averaged to account for deviation. as can be seen below, the longer the series, the less people who actual read it through, this should be a key focus in determining the genuine popularity of a series. this displays a DISADVANTAGE to posting MORE CHAPTERS. As compared to the assumed "advantage" it has gained by being posted more often (same number of people will view the front page per day regardless of how many chapters are present on it). because ultimately there is a lesser completion rate. Yet someone keeps trying to claim that popularity = quality in terms of views, despite the fact that a high quality piece can fail to find readers and a low quality piece can find a fanbase. arguing quality on view count and word count is redundant because it has nothing to do with actual quality.

rank	Chapter View Weight(%)	Chapters(#)	Avg Chapter Views(#)	View Ratio (beginning : recent)	Avg Word Count	Release Frequency (Chapters/week)	Rating	Total Chapter Views
1	75.7%	129	1303	19:1	1269	24	4.1	168126
2	88.9%	340	1845	15.5:1	2331	4	4.3	627539
3	61.2%	21	876	2.6:1	1218	6	4.2	18400
4	63.5%	27	1707	12:1	3932	6	3.7	46095

I agree with @DaoFox with what they said about how the trending tab should be based off the daily views on the latest update of every story, that way it is also fairer to short shorties / oneshots as they are often buried at the very bottom at every list due to their lack of any updates, which is absurd when the site while unofficially hosting short story competitions, still recommends their existence with no support on the actual site besides the front page advertisement during the voting period which is usually a few days to a week, putting multiple chapter stories above these one shots and short stories in rank. It also makes sense since some stories aren't read to completion, or are often dropped mid-way and never read again.

And when I actually think about, most sites don't feature their most viewed books or novels on the front page or as a default search result, it completely washes out the newer stories as it blocks potential readers from bothering to search for lesser known novels. And I'm going to use this analogy in no way as an offense to readers, but simply put, the current trending page being the first thing you see on Scribblehub is like offering a child candy in front of them and the same exact candy a mile away from them, for their own convenience they would just take the candy right in front of them. Why bother searching for a good novel, when the site is offering you some on a silver platter? Sites like AO3, FanFiction, and even Novel Updates, all have their default searching and tag results or their homepage showing the latest updates on novels. If people want to look at highly acclaimed books and novels on the site, they should search for it themselves to make it fairer for everyone on the site, that way no author nor story is prioritized over the other.

So yes, if the trending list is actually able to portray popularity fairly instead of an easily manipulative system like the current (default) daily views, which includes all updates, then it can prioritized on the front page. But right now, the default setting for the trending tab on the front page is based off daily views, which is counted across multiple chapters, with a simple algorithm that is easy to take advantage of. Simply put, the entire trending list being the first thing you see on the site is unfair, it advertises already popular stories over those that aren't.

Speaking of homepages, the site I mentioned above all have different formats for their front pages, but both of the sites I usually use to read fanfiction have a similar set up where they prioritize readers actually searching for the stories they want to read

Similar to these sites, ScribbleHub could utilize a system where the front page actually directs viewers to the genres of the novels, so that readers search for what they want and are in no way influenced by the front page or by a trending list to read something instead. Problem solved, get rid of the trending list if it is too much of a problem or is too controversial. And the site still has the regular search and advanced search option for people who just want to read more popularly viewed stories. Simple as that.

 

[lnv]

The issue seems to be that people are afraid their works aren't being seen, they are looking for ways to "game the system" to get their "much better" novel above people who write a lot more than they do. I say do away with the trending list and put up a random Novel list. That is absolutely fair and when people aren't getting readers they can blame their own writing instead of another author's work or style.

The trending list serves multiple purposes for SH, we can get in those purposes if you wish, but I don't see a problem with just fixing it up to be as humanly fair as possible. Overall, I think featuring on NU gives way more traffic than the trending list and those releasing daily would still have an advantage over those that do not.

But as I pointed out, adding 1 entry or so for randomly featured novel of the day(that fits minimum requirements) may not be a bad idea.

As you can see, there actually is a pretty large advantage to be had in posting multiple shorter chapters per day in terms of the trending list as it is currently designed, and that is not completely fair. The counting of 1 view per unique viewer per day would go such a long way toward leveling this out that I honestly doubt anything else would need to be done. You can continue posting as you like, and if that's what the viewers like then you will still have your story doing well. It will do nothing to change your 24 hour view count on a single chapter, which I think is a far better indicator of the story being liked than total view count given the different factors here.

Is there any reason for unique count per day vs per week? As I mentioned above, some stories like lets say Arcane Emperor which is top of the RRL rankings used to do 2-3 releases per week with 8k-13k chapters. Is there any reason to disadvantage those writing really long chapters but not releasing daily?

the unique views would be ideal, based on chapters and not the series as a whole, but the limitations are in how to achieve a proper representations. IPv4 leaves a lot to be desired and while IPv6 is more accurate, its not perfect. the IP address changes and so do the people using it. it can't be based via users because there are far more readers who see no need to make an account and so we slowly run out of factors to determine the unique views as flaws appear for each method that harm the accuracy and thus fairness.

That isn't really a problem. You have 3 identifiers, IP, session id and account id. You can add digital fingerprint in as well but with the way the net is moving to trying to limit digital fingerprinting, that may not be worth wasting time on. So when a user views a chapter, it will record their ip, session id and account and check for uniqueness.

my only issue with unique views is the "re-readability" factor novels, published or web-alike, possess. when a single persons view counts just once no matter how many times they opted to re-read the series, either because later chapters make them rethink what they read in the early chapters, or because they liked a story enough to refresh their memory, though you would hope this isn't happening due to a necessity for confirming the story-verse's facts and continuity. this isn't a key factor to consider in trending as a whole, but it is an indicator worth taking note of.

Trending is based on most popular per week. So there if we go with unique per week, them clicking again next week would count them again. Thus re-readability isn't hurt.

My own upload being what it was, has 850 views total, only 179 of these are chapter views, bringing my average to 90 views per chapter. I would honestly find it useful to know how many of these are unique in addition to the totals, because it lets me know how many people my work has really reached and how much attention they have paid to it. but this won't be accurate if the same person without an account reads my story multiple times from different locations bringing the validity of the results into question.

I agree that having unique statistics would also be nice regardless of trending or not.

my phone has a different IP than all my other internet capable devices. just viewing it again from one of them will inflate the results and doesn't really indicate that the series is "trending" more so than before. same if my device gets reassigned an IP address between viewings.

Your session id and account would track you across ip changes unless you use incognito/private mode. Device changes, nothing you can do about. But it isn't big enough of a deal. Again, I said as fair as "humanly" possible.

as for the fall off, for example we can take stats from the top trending series as viewed on the homepage. latest view count will be taken from the most recent chapters averaged to account for deviation. as can be seen below, the longer the series, the less people who actual read it through, this should be a key focus in determining the genuine popularity of a series. this displays a DISADVANTAGE to posting MORE CHAPTERS. As compared to the assumed "advantage" it has gained by being posted more often (same number of people will view the front page per day regardless of how many chapters are present on it). because ultimately there is a lesser completion rate. Yet someone keeps trying to claim that popularity = quality in terms of views, despite the fact that a high quality piece can fail to find readers and a low quality piece can find a fanbase. arguing quality on view count and word count is redundant because it has nothing to do with actual quality.

rank	Chapter View Weight(%)	Chapters(#)	Avg Chapter Views(#)	View Ratio (beginning : recent)	Avg Word Count	Release Frequency (Chapters/week)	Rating	Total Chapter Views
1	75.7%	129	1303	19:1	1269	24	4.1	168126
2	88.9%	340	1845	15.5:1	2331	4	4.3	627539
3	61.2%	21	876	2.6:1	1218	6	4.2	18400
4	63.5%	27	1707	12:1	3932	6	3.7	46095

Im confused, is this statement addressing towards me?

Overall, I agree that a higher quality work may fail to find a fanbase while a lower quality one can get a large fanbase. Though I'm not understanding what you mean by ones posting more chapters will be disadvantaged, are you talking about if we switch to uniques?

 

[Jemini]

Is there any reason for unique count per day vs per week? As I mentioned above, some stories like lets say Arcane Emperor which is top of the RRL rankings used to do 2-3 releases per week with 8k-13k chapters. Is there any reason to disadvantage those writing really long chapters but not releasing daily?

Actually, there is, and it's actually to HELP the series that have 2-3 releases per week. If it was on a per-week view count, it would actually favor the series that release more chapters per week a lot more since they have more NU exposure.

If it's per day, the ones that release 2-3 per week will get their exposure on the day they release, and be able to spend a day on the trending page if they are good enough for it. Then, others can fight their way to the top on other days. Making it on a per-day basis will actually vary up results on the trending page a lot more instead of getting the same results every day.

If it was on a per week basis, not much would change for the ones we see at the top from what we're seeing now. Plus, I imagine a per day tracking would also be easier to implement.

 

[lnv]

Actually, there is, and it's actually to HELP the series that have 2-3 releases per week. If it was on a per-week view count, it would actually favor the series that release more chapters per week a lot more since they have more NU exposure.

If it's per day, the ones that release 2-3 per week will get their exposure on the day they release, and be able to spend a day on the trending page if they are good enough for it. Then, others can fight their way to the top on other days. Making it on a per-day basis will actually vary up results on the trending page a lot more instead of getting the same results every day.

If it was on a per week basis, not much would change for the ones we see at the top from what we're seeing now. Plus, I imagine a per day tracking would also be easier to implement.

Uhm, trending is based on per week as is, so if you do it per day, if you release 2-3 per week, you would count as 0+1+0+1+0+0+1/7=0.43 vs someone who releases every day would be 1+1+1+1+1+1+1/7 = 1.

If you are talking about also changing trending to be based on daily basis, then wouldn't those who released their story at 12 oclock have an advantage over those that release mid day? Or do you mean it'll base off the previous day?

I don't see any disadvantage for 2-3 releases in weekly because it is unique views. So if 50 people read 10 chapters per week or 1 chapter per week, it won't make a difference. The only advantages those that release every day have is the general advantage of frequent featuring on NU and always being on the front page results. But if you also want to change it to change daily, then how would you go about that?

 

[Jemini]

Uhm, trending is based on per week as is, so if you do it per day, if you release 2-3 per week, you would count as 0+1+0+1+0+0+1/7=0.43 vs someone who releases every day would be 1+1+1+1+1+1+1/7 = 1.

If you are talking about also changing trending to be based on daily basis, then wouldn't those who released their story at 12 oclock have an advantage over those that release mid day? Or do you mean it'll base off the previous day?

I don't see any disadvantage for 2-3 releases in weekly because it is unique views. So if 50 people read 10 chapters per week or 1 chapter per week, it won't make a difference. The only advantages those that release every day have is the general advantage of frequent featuring on NU and always being on the front page results. But if you also want to change it to change daily, then how would you go about that?

Err... no, trending is based on per-day views as is. Strange, I do not know where the confusion here is coming from. At any rate, weekly trending lists would still be a lot more static. It should be best to have it alter more frequently in order to give more people a chance, thus daily sounds like the more fair model.

 

[Phantomheart]

Err... no, trending is based on per-day views as is. Strange, I do not know where the confusion here is coming from. At any rate, weekly trending lists would still be a lot more static. It should be best to have it alter more frequently in order to give more people a chance, thus daily sounds like the more fair model.

Uhm, trending is based on per week as is, so if you do it per day, if you release 2-3 per week, you would count as 0+1+0+1+0+0+1/7=0.43 vs someone who releases every day would be 1+1+1+1+1+1+1/7 = 1.

If you are talking about also changing trending to be based on daily basis, then wouldn't those who released their story at 12 oclock have an advantage over those that release mid day? Or do you mean it'll base off the previous day?

I don't see any disadvantage for 2-3 releases in weekly because it is unique views. So if 50 people read 10 chapters per week or 1 chapter per week, it won't make a difference. The only advantages those that release every day have is the general advantage of frequent featuring on NU and always being on the front page results. But if you also want to change it to change daily, then how would you go about that?

Yo, trending is based on many things, it has a couple of options here guys, but the default trending that you see on the homepage is indeed based off DAILY VIEWS.

 

[lnv]

Err... no, trending is based on per-day views as is. Strange, I do not know where the confusion here is coming from. At any rate, weekly trending lists would still be a lot more static. It should be best to have it alter more frequently in order to give more people a chance, thus daily sounds like the more fair model.

Yo, trending is based on many things, it has a couple of options here guys, but the default trending that you see on the homepage is indeed based off DAILY VIEWS.View attachment 952

Sorry, my mistake. I could have sworn it used to be weekly in the past... either that or I came from alternate universe where it was weekly...

Well that clears up my confusion there.

 

[Phantomheart]

Still don’t understand why we have a trending list anyway, most writing and novel sites don’t use them unless they are used for promoted stories. Most novel hubs leave it up to the reader to search up for what they want to read usually. And then people who like the same novels recommend similar novels. Trending is just some of the different filters mashed into a single tab called trending, it doesn’t make sense to juxtapose the most viewed novels right next to the least viewed novels in sight if you are trying to give equal standing in all stories, I understand putting latest updates on the homepage, but why the most viewed? Doesn’t make sense, especially since ranks are already a thing.
Edit: only site that I know of that has the most trending things on their homepage would be AnimePlanet which is more of a review site if not anything else.

 

[SealJohnson]

Still don’t understand why we have a trending list anyway, most writing and novel sites don’t use them unless they are used for promoted stories. Most novel hubs leave it up to the reader to search up for what they want to read usually. And then people who like the same novels recommend similar novels. Trending is just some of the different filters mashed into a single tab called trending, it doesn’t make sense to juxtapose the most viewed novels right next to the least viewed novels in sight if you are trying to give equal standing in all stories, I understand putting latest updates on the homepage, but why the most viewed? Doesn’t make sense, especially since ranks are already a thing.
Edit: only site that I know of that has the most trending things on their homepage would be AnimePlanet which is more of a review site if not anything else.

Royal Road has a trending list that's based on percentage increase of total ratings beginning 30 days after a series crosses a certain rating threshold. That is: (new ratings) / (ratings 30 days ago).

New series with low numbers of ratings can easily double that number and earn a spot on the list. Series that already have tons of ratings have to get ever increasing numbers of ratings to stay on the list. As a result, it's essentially impossible for even the most wildly popular series to stay on the list for more than a month or two.

It's not a perfect system, but it does draw eyes to popular new series and hidden gems a lot more successfully than on most other sites, including this one.

Of course, the trending list is only part of the problem. Just as important are the fact that all the rankings and all the default lists generated by the search tools sort based on the same almost-meaningless stat. The 'reader' count gives roughly the same popularity measure without the chapter length bias, and is less susceptible to cheating. By cheating I don't just mean writing shorter chapters. Authors can also just ask a few friends to flick through chapters over and over to artificially increase views. Views are barely better than the roll of a dice for gauging actual popularity.