Thanks everyone! As I said, this is a very simple recommendation engine. Basically, all it's doing is recommending the works that your friends' like. So, it's not going to be terribly accurate when taking into account your own tastes. Ideally, what it should do is take all the likes you've ever given and then compare that against everyone. That way, if your likes are, say, 90% similar to X (whether or not you've even met X before), then the recommendation engine would recommend you something X liked but you haven't read yet.
That's possible, and I have a rough idea of how to do it, but it'd require lots of time and effort. Moreover, I'm not even sure if it'd be an improvement on the current system. For one, just because you and X like 90% of the same material, that doesn't mean you'll like everything X likes. For another, part of the fun of the current one is seeing what your YWS friends are reading. Even if the work isn't to one's taste, I think we all still want to read what our friends are reading.
Still, I am adding one improvement, and maybe two. The first improvement is it won't recommend stuff that you've commented on but haven't liked (currently, it'll recommend it even if you've already commented on it). Unfortunately, I have no idea whatsoever how to keep it from recommending stuff you've viewed. Somewhere, the YWS database keeps track of stuff you've viewed, but I have literally no idea how it does that.
The second possible improvement is to add more points for stuff your friends have not only liked, but commented on as well. After all, if someone likes and comments on something, then it must be better than something they only liked. I say possible improvement because this could cause unwieldy page loading times. If it does, then I won't add this improvement (page-load times are much more important).
Gender:
Points: 1193
Reviews: 262