How Canonova Score Works

Canonova Score is a dynamic ranking system that evolves with every decision you make. Unlike static star ratings that remain fixed once assigned, Canonova Score adapts in real-time as you compare movies head-to-head, creating a living, breathing hierarchy of your true preferences.

The system is inspired by ranking methods used in competitive gaming and sports, adapted specifically for movie preferences. It recognizes that your feelings about movies aren't static—they change based on context, mood, and new experiences. Canonova Score captures this evolution, ensuring your rankings always reflect your current preferences.

What makes Canonova Score unique is its focus on relative preferences rather than absolute ratings. Instead of asking "How much do I like this movie?" (which can vary day to day), it asks "Which movie do I prefer?" (which is easier to answer consistently). This approach produces more accurate and stable rankings than traditional rating systems.

When you first import your movie collection from Letterboxd or IMDb, each movie receives an initial Canonova Score based on your existing star rating. This gives your movies a head start that reflects your current preferences. Movies you've rated highly start with higher scores, while lower-rated movies begin with lower scores.

This initial scoring ensures that your Canonova rankings immediately reflect the work you've already put into rating movies on other platforms. You don't start from scratch—your existing ratings provide the foundation. However, these initial scores are just the beginning. As you make comparisons, scores adjust to reflect your true preferences, which may differ from your original star ratings.

For example, you might have rated two movies 4.5 stars on Letterboxd, but when forced to choose between them, you consistently prefer one. Canonova Score will recognize this pattern and adjust the scores accordingly, creating a more nuanced ranking than star ratings alone can provide.

The heart of Canonova Score is the comparison system. Every time you choose one movie over another, both movies' scores adjust. The movie you prefer gains points, while the other loses points. This creates a self-correcting system where movies naturally find their true position in your personal canon.

The amount of points exchanged depends on the current score difference between the two movies. If you choose an underdog over a favorite, the underdog gains more points, reflecting the significance of that decision. Conversely, if you choose a highly-ranked movie over a lower-ranked one, the point exchange is smaller, since this outcome was more expected.

This dynamic adjustment means that your rankings become more accurate over time. Movies you consistently prefer rise to the top, while those you rarely choose settle into lower positions. The system accounts for the context of each comparison, ensuring that a close call between two great movies is weighted differently than an easy choice between a favorite and a movie you barely remember.

The system intelligently selects movie pairs for comparison, prioritizing matches that will be most informative. It focuses on movies with similar scores, ties, or movies that need more data to find their true position. This ensures that every comparison you make contributes meaningfully to building your accurate ranking.

Canonova Score organizes your movies into 10 distinct tiers, each representing a different level of preference. This tier system provides a clear visual hierarchy that makes it easy to understand where each movie stands in your personal canon.

Tier 1 (2300+ points) represents your absolute favorites—the films you'd take to a desert island, the ones you'd show to someone who wants to understand your taste. These are the movies that consistently win comparisons and have proven themselves as your top choices.

Tier 2 (2100-2299 points) contains excellent films that you love, but perhaps not quite as much as your Tier 1 favorites. Tier 3 (1900-2099 points) includes very good films that you enjoy and appreciate. As you move down through Tiers 4-10, you'll find movies with decreasing scores, down to Tier 10 (0-699 points), which contains films that rarely win comparisons.

The tier system is dynamic—movies move between tiers as their scores change. A movie that starts in Tier 3 might climb to Tier 2 as you make more comparisons and discover you prefer it more than you initially thought. Conversely, a movie in Tier 1 might drop if you consistently choose other movies over it in new comparisons.

This fluid organization reflects the reality that your preferences evolve. The tier system captures this evolution, giving you a clear picture of how your movie canon is developing over time.

The key to building an accurate Canonova Score ranking is making honest comparisons. Don't overthink your choices—go with your gut feeling. The system is designed to self-correct over time, so individual comparisons don't need to be perfect. What matters is consistency in your preferences.

When comparing movies, consider which one you'd rather watch right now, which one you'd recommend to a friend, or which one you'd choose if you could only watch one. These questions help you access your true preferences rather than what you think you should prefer.

It's normal for your choices to vary based on mood or context. The system accounts for this variability—it looks for patterns across many comparisons rather than relying on any single decision. Movies that consistently win comparisons will naturally rise, while those that rarely win will settle into lower positions.

Many users find it helpful to do 10-20 comparisons per session, gradually building up their canon over time. You don't need to compare every movie against every other movie—the system intelligently selects pairs that will be most informative. Focus on making honest choices, and let the algorithm handle the rest.

Remember that there are no "wrong" choices. Your preferences are personal and valid. The goal isn't to create a ranking that matches critical consensus or popular opinion—it's to build a ranking that accurately reflects your unique taste in movies.

Canonova Score offers several advantages over traditional star rating systems. First, it creates a true ranking rather than a collection of independent ratings. With star ratings, you might have dozens of movies at 4.5 stars, but no way to distinguish which you actually prefer. Canonova Score solves this by creating a precise hierarchy.

Second, Canonova Score is dynamic and adaptive. Your rankings evolve as you make comparisons, reflecting changes in your taste over time. A movie you loved years ago might settle into a lower tier as your preferences mature, while new discoveries might rocket to Tier 1. This evolution is captured automatically, without requiring you to go back and update hundreds of old ratings.

Third, Canonova Score accounts for the relationships between movies. It recognizes that your preference for Movie A over Movie B is meaningful information, even if both have the same star rating. This relational data creates a more accurate ranking than absolute ratings alone.

Finally, Canonova Score is self-correcting. If you make a comparison that doesn't reflect your true preference (perhaps due to mood or context), subsequent comparisons will correct the ranking. The system looks for patterns across many decisions, ensuring that temporary variations don't permanently skew your rankings.

Your Canonova Scores are useful for more than just ranking movies. They help you discover hidden gems in your collection, identify patterns in your taste, and create meaningful lists to share with others.

The tier system makes it easy to explore your collection. Filter by genre, decade, or specific year to see how your preferences vary across different categories. You might discover that your Tier 1 movies are dominated by a particular genre or era, revealing something about your cinematic taste.

Canonova Scores also help you make better viewing decisions. When you're looking for something to watch, check your Tier 1 movies—these are your absolute favorites, guaranteed to satisfy. Or explore Tier 2 and Tier 3 to find excellent films you might have forgotten about.

The Top X feature lets you create beautiful, shareable images of your rankings. Generate lists of your top 10 movies, your favorite sci-fi films from the 2010s, or your entire Tier 1 collection. These images are perfect for sparking discussions, discovering new films, and showcasing your personal movie canon.

You can also publish a read-only public link to your live tiers from the Tiers page (Publish Link). Others can browse your Tier 1–10 layout without editing anything. Choose a static snapshot or automatic updates, and unpublish whenever you want. Your collection stays private until you opt in. Community ranking shares for individual titles work the same way: private until you share.

As you continue making comparisons, your Canonova Scores will become more accurate and stable. The more data the system has, the better it can distinguish between movies and create a ranking that truly reflects your preferences. This is why many users find their rankings become more meaningful after they've made hundreds of comparisons.

The Export page (user menu → Export, or from Profile) downloads your collection as a CSV. You can take a full backup, pull scores-only data, or build a file for Letterboxd's importer.

Presets set the columns for you: Full library for a complete Canonova dump, Letterboxd import for Title / Year / IDs / Rating / Review in Letterboxd's header style, and Scores only for Canonova Score, overall rank, tier, and related fields. You can also tick individual fields for a custom mix.

Optional filters (genre, decade, year, media type) and sort options limit which titles go into the file. The Letterboxd preset's Rating column maps your Canonova tier to a 0.5–5 star scale, and Review uses the same Movie Critic–style text as sharing (Canonova Score, overall rank, stars, and decade/year/genre ranks). Ranks in that text use your full collection even when the export is filtered.

Importing that CSV on Letterboxd can overwrite existing Letterboxd ratings for matching titles—review the file before you upload it there.

From a movie's details (for example on Search or Tiers), Share as Letterboxd Review builds the same critic-style text used in Export's Review column. Letterboxd's website does not let other sites pre-fill the review box, so Canonova walks you through a short copy-and-paste flow.

Tap the button to see the review text, then 1. Copy review to clipboard and 2. Open Letterboxd and paste review. Open uses the same Letterboxd film as the small Letterboxd icon on the modal (including resolving short links to the correct title's review page when possible). On iPhone, you may also open the Letterboxd app, which can prefill the review field.

You can still Share Ranking with Community on the same modal to publish a single-title ranking for other Canonova users, separate from posting on Letterboxd.

The Games page (in the main navigation, between Search and Stats) lets you test your movie knowledge with two collection-based games. Both use only movies you've uploaded, so they're fully personalized.

My Movie Trivia uses AI to generate 10 multiple-choice questions from your collection—about directors, actors, release years, and more. After you finish, you see your score and can click "Play 10 New Questions" to start a new round; the page scrolls to the top so you're ready for the next set.

Keyword Quiz picks a random movie from your collection and reveals its keywords one at a time. You search your collection by title (same search as on the Search page) and pick the movie you think it is. If you run out of keywords, you can reveal extra hints in order: the year, then the director(s), then the actors. A "Give Up" button shows the correct answer and the movie poster. When you get it right or give up, the guess area is replaced by the answer so you see the result clearly.

Games are a fun way to engage with your canon and see how well you know your own rankings and metadata.