A movie rating predictor answers one narrow question: what would you rate this film? Not the crowd, not a critic. A handful of apps try to answer it, and they do it in two very different ways. This guide compares them as of October 2026, including where each one is the better choice. Kinolog is one of them, and we make it, so judge our section with that in mind.
Two ways to predict a rating
People like you. Collaborative filtering finds users whose past ratings agree with yours, then predicts your score for a film from how they rated it. It needs a large community and works best for popular films many people have rated. Criticker, MovieLens and Taste all start here.
Your own history. A diary-based predictor compares a film with the films you already rated: same director, same writer, the same kind of film, the same era. It works for obscure films as long as your diary has related ones, and it can show its working, because the evidence is your own ratings rather than strangers’.
Neither approach is magic. Both are wrong sometimes, and the honest ones say how often.
The apps at a glance
| App | Predicts from | What to know |
|---|---|---|
| Criticker | Other users with similar taste | Web. Rate on a 0 to 100 scale; scores are compared as percentiles, and it matches you with like-minded users before it predicts. |
| MovieLens | Other users with similar taste | Web. A non-commercial research recommender run by the GroupLens lab at the University of Minnesota since 1997. |
| Taste | Other users with similar taste | iPhone and Android. Rate films awful, meh, good or amazing and it shows a Match percentage; it suggests rating about twenty films first. |
| Kinolog | Your own diary | iPhone and web. A private movie diary that predicts the star rating you would give, names the films behind it and checks itself later. Predicted stars are part of Plus. |
| Letterboxd | Not offered | iPhone, Android and web. The best place to log films with friends; it does not give a personal predicted rating before you watch. |
Which one should you use?
- You love numbers and the web: Criticker. Its 0 to 100 scale and taste-matching have a loyal following, and it has been doing this for years.
- You are curious how recommenders work: MovieLens. It is a free research project, and its public rating datasets are among the most used in academic work on recommendations.
- You want something quick on your phone: Taste. Four-word ratings are the fastest way to give an app something to work with.
- You keep a diary and want to see the reasons: Kinolog. It imports a Letterboxd history in one step, predicts from your own ratings, and refuses to predict at all until it has enough evidence.
- You mainly want friends and reviews: stay on Letterboxd. None of the above replaces its community.
How Kinolog’s prediction works
Kinolog starts from how you use the star scale, then looks for evidence among the films you rated: who directed it, who wrote it, the specific kind of film, its genre and its era. It names the closest films it compared against, labels the call as a hunch, a read or confident, and refuses to predict at all below eight rated films. The number is computed from your diary, not written by a language model.
Every prediction is saved when it is made and settled against the rating you give after watching, so you can see your own hit rate. Kinolog does not publish a product-wide accuracy figure yet, because the settled sample is still too small for one to be honest. The full method is in the prediction methodology, and the rating predictor page shows what a call looks like.
How to judge any rating predictor
- Does it say when it does not know? A prediction from five ratings should look less certain than one from five hundred.
- Does it use your scale? A generous rater and a harsh one can mean the same thing with different numbers.
- Does it keep score? Record the prediction before you watch, then compare. A predictor that only shows its hits is marketing.
- Can you leave? Your ratings are years of work. Pick an app that exports them.
For the difference between a personal prediction and a crowd score, read personal movie rating predictions vs audience scores.