FF Team:
Just wanted to respond to this.
The recommended search, which is the default search until you change your search parameters, is actually in favour of well written stories. We have designed it this way to promote good quality stories so they can get to the top of the listings.
There are several things that factor in when the recommended search displays.
* Number of likes: This is factored in because members will like content that is well written.
*Number of chapters: Longer stories mean that the author has taken the time to develop the characters and plot making It a more interesting read.
*Thumbnail and description: Again this is about the time and care spent on the story. If the author can be bothered to fill these sections out correctly we hope that time and care has been taken on the main text itself.
* Newness:This is also considered so that there is a rotation of stories. That are still of good quality... as it includes the other points above too. We don't want a stagnant page that only shows the top stories that people have already read. There are always new and upcoming authors that should be showcased!
Of course everyone of our members apply different qualities to stories that make them interesting to them personally. Maybe it's the writing style, a particular author your following, a certain scenario may take your fancy or a particular character role. Whatever it is, you can use the search function to find stories that fit your preference.
We try to keep things diverse and don't want the stories section to become stagnant. Obviously algorithms are never perfect but there has been a lot of thought put into how this works. However, if you do have any bright idea's on how to improve this section, please send them our way!
You could track and network 'like' statistics as other social sites sometimes do. Each reader has a preference profile, but it isn't often easy to quantify. So don't quantify it. Instead, go top-down: For instance, if one user has liked 30 different stories, and a new user has liked the 5 of those that they've seen, there's a probability that the other 25 might be ones they would like as well. This is a very simple example, of course, but in the modern era using something as basic as just 'likes' doesn't really cut it anymore. You need to be understanding and categorizing the types of likes with respect to the users giving them - namely that different users like different types of stories, and generally have identifiable preferences - and then using that information to recommend content and, ideally, create preference profiles to further help develop your algorithms.
