Right now, it is seeming to me that people are bookmarking chapters and favoriting stories rather than liking them so as to avoid giving a leg up to what they see as the 'competition' for things like "Story of the Week" and the like. I propose that we quash this sort of anti-social behavior by adjusting the algorithms that pick SotW and all other top ranks to factor in ALL tracked data metrics for each chapter and story: likes, favorites, bookmarks, views, comments; the works. Obviously, this would be a major adjustment and would need quite a bit of black-box beta-testing and tweaks to avoid abuse before rollout, but I think it's necessary to preserve the inclusive and community-minded atmosphere that presently exists on this site which currently makes it such a great place to write and prevent it descending into the kind of cliquish and vicious cess-pit that places like Annon-Land have become...
I myself have some frustration with bookmarks/favorites not being considered in the stories score but is this becoming s problem here? I know downvotes we’re used in the past to nuke people and had to be done away with but is there any evidence now of this being used in a similar fashion.
Fortunately, it CANNOT be used in a particularly similar fashion, as it cannot negatively impact peoples scores, only avoid positively impacting their scores. However, that second instance is exactly how I see it being used.
Why are people so petty that they want to manipulate scores to achieve something as minor as story of the week? Besides if you have to cheat to do it you what did you really win?
Why should the "number" of comments influence the rating? If there are a lot of comments, it could be a quest story or comments to point out typos. That doesn't say anything about the quality of a story. Stories and chapters on the front page should get more views as they generally get more attention. I also think that there would be more bot activity. That would just highlight the stories which are already in the "sweet spots." Favorites and bookmarks only hint at how many readers are actually reading a story. Once again, popular stories in the "sweet spots" are more likely to gain favs (and bookmarks). What about if there would be a stat "average likes per chapter per favorite" which would influence the ranking. (excluding bookmarks) To include that stat for ranking purposes, a certain number of favorites could/should be necessary (like 50). The "average likes per chapter" would show the average quality of the whole story. Maybe deeper chapters could get some advantage, though I think that would be difficult to implement (resource expensive). Now, if you divide the "average likes per chapter" by the number of favorites, you'll get the "percentage" of readers out of the pool of potential readers (those who faved the story) who liked a chapter. Example... Story 1 Chapters: 40 Likes total: 240 Favorites: 50 -> Average likes per chapter: 6 (240/40) Average likes per chapter per fav: 0.12 (6/50) -> Percentage: 12% (0.12 x 100) Story 2 Chapters: 40 Likes total: 240 Favorites: 100 -> Average likes per chapter: 6 (240/40) Average likes per chapter per fav: 0.06 (6/100) -> Percentage: 6% (0.06 x 100) I think that would give good stories which aren't as well-known the rating they deserve.
That makes sense for views, but but for favorites and bookmarks, they're no more likely than likes to be only how many people see it. Averaging likes per chapter by "potential readers" rather than by actual readers is pointless when that's also a statistic we have for a chapter. Also, one of Nemo's proposed problem was that people are favoriting stories rather than liking them (I don't know if that's true myself), and if it is, your system would make it so that favoriting a story, saying that you want to read more of it, but not liking any chapters, would be even worse. If it's just likes, than those two stories you mentioned are tied, but your system scores the second one lower for having more people favorite it, meaning that authors would want people not to favorite their stories if possible.
@gene.sis Yes, your system turns the favorite into essentially a new version of the dislike button, which is the exact opposite of what its sposed to be. EDIT: there's also a typo in your math.
“I don’t think it’s as cut and dry as favorites are the new dislike. Dislikes had strong negative connotations. Favorites are slightly positive as they do indicate story quality. People favoriting and not giving likes and only favoriting or bookmarking has been a point of consternation for quite some time. I just thought there might be something where people have agreed to only like certain peoples chapters. But I don’t know if this assumption of yours rises to more than a conspiracy theory.
How dare these people to only favorite a story, and not hit the like button on each and every one of its chapters?!? If it weren't for the audacity and insolence of these people, every one of our stories would be ranked in the top 10! They are like locusts, feeding on our creativity, taking everything, but giving nothing back in return. Don't they know how much pain they cause us, when they interrupt our daily fix of like-notifications with something entirely meaningless as a bookmark? If we let them go on with their shameful behavior, they will destroy everything that we hold dear! So I say that we shall not suffer them any longer! I urge the site administration to finally take a stand and demand that they take drastic action to exterminate the scourge that these people represent. First time offenders should be publicly shamed, while repeat offenders should have their accounts and ips banned immediately, to prevent them from every polluting our cozy community again. That way we will be save once again, and can enjoy our daily fix of likes without getting bothered by undesirable notifications of people who dare to hit the favorite @Nemo of Utopia: I'm all for using additional metrics, i.e. favorites, bookmarks, the ratio of chapters to likes, etc. But i think that your opening statement is a bit far fetched. How did you come to that conclusion, that people "not like" on purpose to give other stories an edge in the rankings? @gene.sis: I dont't get it, why should a story with identical stats (besides the favorites) be rated worse than a story who has less favorites?
While the satire stings a bit, it does so because there's a ring of truth to it. I have no direct evidence that they are I've just been noticing favorites increasing while likes have dropped off. I probably jumped the gun with this one.
I've found this too. Seems like a heap of favourites but very few likes coming in. Bookmarks I think are used as intended, at the end of a string of chapters as a placeholder until more are written.
Not sure if I understand that correctly... More Favorites doesn't necessarily mean a better story. It could just be older or got more attention. Potential readers: all CHYOA users Actual readers: all users who faved the story (not all, but it's likely that it is always about the same share; readers who like chapter beyond the introduction seem to be really rare) So currently we have "Likes per potential readers" "Likes per Favorite" would be much more near "Likes per actual reader" In my experience, there are nearly no users who like a chapter without favoriting the story. I think only about 20% of all readers use the like button. The other 80% don't use it for whatever reason. If 50% of the like-button-users like a chapter, it is pretty likely that also 50% of the like-button-refuseniks like the story/would press the button if they would use it. Likewise, if there are 20% of the one, there would also be 20% of the other. About the difference between favoriting and liking... all the readers who fav a story and read its chapters but don't use the like button have an opinion about the story, even if they don't (want to) express it. These opinions will most likely pretty similar to the opinion of the users who liked the chapter, so it will make no statistical difference if you project the actual likes to the actual readers. Let's reverse the example: Both of our example stories have 40 chapters and 100 favorites. Now Story 1 has 480 likes while Story 2 has 240 likes. So I'd assume Story 1 has a better quality as there is a higher percentage of readers who like it. corrected it, thanks (Damn copying ^^) The difference is that a "malicious favorite" would affect the higher value. A story which had 100 Favs, 10 Likes pc and 0 Dislikes pc had a Rating of 100% A change to 1 Dislikes pc changes the Rating to 90.9% (a reduction of 9.1%; 10 Dislikes lead to a reduction of 50%) In case of an added favorite, that rating would change from 10% to 9.9% ( a reduction of 1.0%; 10 added Favs would lead to a reduction of 9.09%) But it could indeed be affected by malicious activities, though it doesn't have to be the only thing of an overall ranking algorithm. That is the problem with favorites... while that number is indeed affected by the quality of the story, there are other aspects which raise that number like chapter depth, the number of chapters, visibility on the front page and the time when a story is visible on the front page. Basically, because there are more people who read it, so obviously a lower percentage of them liked it. Does the reversed example above make it more clear? Imagine two identical flowers. One of them is located downtown, the other can be found deep in the woods. During the life of these flowers, they achieve following "stats": Downtown flower: Noticed by 1,000 people, enjoyed by 500 people Flower in the woods: Noticed by 100 people, enjoyed by 50 people So which one of them is nicer? They are both equal, aren't they? Every second person who noticed each of them enjoyed their sight. Now, let's take another set of flowers Downtown flower: Noticed by 1,000 people, enjoyed by 50 people Flower in the woods: Noticed by 100 people, enjoyed by 50 people Which of them is nicer? Both were enjoyed by the same amount of people... The thing is, the downtown flower just got more attention and even if it isn't as nice as the flower in the woods, the same number of people enjoyed its sight.
Okay, but when you look at a chapter, you can see how many people have read it, so you don't need to use favorites to "guess" how many people read it. I'll cop to forgetting to like chapters all the time. If I read a story, and most of the time, when I reach the end of a chapter, I click to the next one without liking it because I want to keep going, and then when I reach the end and there's no more, favorite it so that I can get more of it in the future when more chapters are written, you're saying that shows me to not really like the story that much. Also in your example What if we look at the number of views on the first chapter, and see that Story 1 has 2000 views, and Story 2 has 500 views. Now we have more information. Of the 500 people who saw Story 2, 1/5 favorited it. of the 2000 people who saw Story 1, only 1/20 favorited it. Now saying that Story 1 is better because it has more likes per favorite doesn't seem so reasonable.
Views are highly unreliable and don't say much about how many people actually read it. - Accessing a chapter whether you're logged in or not adds 1 to the view count. If you refresh your browser page adds 1 to the view count. Doing that automatically could be used to quickly manipulate that number. - The views from bots get counted as well (at least I haven't seen anything in the code yet which would try to prevent that.) They access the site from different starting points and links (like the forum). Chapters/stories which are present on the main page are more likely to be accessed more often than others. I'm not sure if a bot will trigger the loading of the chapters above 1,000 on a story map so that chapters might have lower views because of that. - If a chapter gets some comments, all the commenters get notified about the recent comment (depending on settings). Let's say a chapter got 10 comments with a time gap of a day from 10 different users, you could have 1+2+3+4+5 = 15 additional views after 5 comments and 15+6+7+8+9+10 = 55 additional views after 10 comments. Comments also attract clicks on the chapter (I occasionally do that if there are some comments on a chapter because "What do they have to say?") - When reading, I usually go back one chapter and start with the last one or two paragraphs to remember what had happened before I switch to the new chapter. This applies to chapters which are published with time gaps of a few days or weeks but not if someone reads a whole branch. - When writing, I reread several parts of the story to be sure that the information in my head is accurate (like "Her panties were red, weren't they?") - Game mode chapters and branching chapters will receive many more views than chapters of kind of linear stories. That can easily happen when Story 1 got published chapter by chapter over a few weeks while Story 2 was published all at once. That should be especially true for the Introduction chapter. So it doesn't add any real information as they aren't from unique internet users (like explained above.) It seems like you're doing that consistently, so statistically, there is no problem as you would probably match a reader (set of readers) who push the like button if you would do but don't push it you wouldn't. Though you could try to make it a habit to think of pushing the button or not when you reach the end of a chapter
Ok, I'm going to say this now, though I doubt it will change anyone's mind: my basic Idea with this suggestion was that you essentially couldn't interact with a story in ANY WAY as a reader without improving its ranking somehow: your only option to not increase the global rank of a story would be to simply not read it, or if you had already read some of it and decided you disliked it to leave and not come back. As others have pointed out, bot-clicks would have to be filtered, I mention this kind of problem in the first post: And as others have said, the stories on the front page would become even more static than they already are, but this has inspired a new idea which I'm going to put up for consideration in a minute here. In any case, if the concern is Readers V Likes it would be the simplest of matters to track how many people visit a story, (is there a chapter which is neither green nor blue for that user's story map? Then they've visited the story,) and use THAT, fully accurate, statistic to judge average likes per chapter vs total readers...
it's not just bots... you'd also have to make sure that only unique users are counted, which means logged in users. (which is basically the same as your suggestion below on a per chapter basis) Doing that on a per chapter basis would indeed result in a perfect stat.
