We structured the recommendations to be 100% personalized, so it might be that you have more a popular taste?
If not please let me know as it should really recommend long tail content too. Could also be that our rated item set is still not very large.
We just use FB Likes to let you rate those (to increase relevance for items we ask you to rate) but don't use FB Likes in the recommendations.
Yeah def. gonna be some movies in your watchlist that you've seen already, we hope it creates trust in the recommendations and we made rating super fast - one person rated 1,600 movies in 90 mins :)
Thanks for the feedback.
The rated item base for now is not yet large enough to give good recommendations to a large group of people.
We'll be able to extend group size in the future.
We've been working on a custom engineered collaborative filtering algorithm not only to recommend movies for one person but for several to watch together, at the same time.
The site is written in Java, using Tapestry 5 & jQuery. Persistence via MongoDB.
We're still in closed beta but would love to get feedback before launching.