Personalized movie recommendations based on what you've liked or disliked.
This free app will accurately predict what you should watch or skip. You can rate any movies and series you've seen in your life. It'll analyze your ratings to understand your tastes and to predict which rating you would give to the shows you haven't watched.
This kind of system that can predict your ratings was made famous by Netflix in the 2000s. Their Cinematch algorithm was very accurate, and they improved it even further in 2006 thanks to a 1 million dollar challenge called the Netflix Prize
But in 2017, Netflix abruptly changed its recommendation system, unfortunately for the worse. They replaced their star rating system with a less nuanced thumbs rating system, but their worst change was that they stopped predicting your ratings to give you instead a percent match score
which, in their words, "indicates how likely we think you are to like that title."
A higher percent match score doesn't mean you'll like more, it means you're more likely to like. They've replaced a quantity with a probability. Netflix no longer predicts if you'll love, if you'll only like, or if you'll hate. Instead, a 70% match score means there's a 70% chance you'll like... or maybe even love... but there's still a 30% chance you'll hate! What kind of information is that?
The final straw: Netflix hides the percentages below 50% as they would indicate that you're more likely to dislike than to like. They willfully stopped warning you when they know you're unlikely to enjoy a show. Is it a coincidence it happens now that Netflix produces most of its content?...
Our recommendation engine predicts HOW MUCH you'll like a show.
Coollector Movie Database predicts your ratings using collaborative filtering
, the same technique that Netflix used. It locates users whose tastes and interests are the most similar to yours. If those like-minded users have seen movies that you haven't, it can assume that you'll enjoy them as much as they did.
The first predictions won't be highly reliable because the algorithm still doesn't have enough information about your tastes to find the best like-minded users, which will quickly improve as you'll rate more movies. The reliability can also vary from one prediction to another because some movies are more popular than others and receive many more ratings. As a result, some predictions are safe bets, but others are more riskier and could be wrong. The app is open about it and displays the reliability of each prediction. The reliability increases as you give the app more clues. The higher the reliability, the more surprising it would be that the prediction was wrong.
You won't find a better movie recommendation system.
The recommendations are adjusted in real-time
, as soon as you rate a new movie. Our engine is even more accurate
than Netflix ever was, because despite having great engineers, it had a big weakness: their limited catalog. If your all-time favorite movies were missing from their catalog, you could not rate them. They missed out on a ton of information about you, in contrast with our algorithm, which has the opportunity to reach a much deeper understanding of your tastes.
Last but not least, our recommendations are entirely unbiased
, contrary to many corporations who tend to promote some content above the rest for obvious business reasons. Maybe you've had some bad experiences before with user-hostile recommendation systems, and you don't trust the algorithms anymore, but please give our system a try. Hopefully it'll change your mind.
Coollector Movie Database is a FREE Mac/Windows software
. No crippled features nor time limits.
There's only a 30 seconds window on startup that you can remove by becoming a contributor.