Rank your anime list with pairwise comparisons.
Pick which item is better until you're satisfied with the order, then export or upload the results.
Import
Connect AniList or upload a CSV.
Filter
Optionally narrow down by rank or year.
Compare
Pick which item is better.
Export
Upload to AniList or download as CSV/JSON.
How it works
Uses a statistical model to learn your preferences from pairwise comparisons. Prioritizes asking about items you're uncertain about, so you don't waste time on obvious choices.
If you're rescoring an existing list, your current scores are used as a starting point.
Technical details
Fits a Plackett-Luce preference model using the MM algorithm. For rescoring, original scores seed the model as a soft prior (MAP regularization). Pair selection uses active-learning policies. Uncertainty via bootstrap resampling. Tier list uses 1D k-means. Inspired by Gwern's Resorter.