Each mode ranks my scores and the external scores among exactly the same films. “All three” includes films with scores from IMDb, Rotten Tomatoes and Metacritic. A single-source mode uses every film with a score from that source. My comparison percentile can therefore change when I switch sets.

Tied scores share their average rank. Percentiles run from 0 to 100; a set with one film has no meaningful percentile. Search, genre, decade and grade filters narrow the displayed ledger without recalculating its percentiles. My existing global archive percentile appears in the expanded details.

A gap is my percentile minus the source percentile, in percentile points. Agreement is 100 minus the absolute gap. With all three sources, the index uses their mean absolute gap, so opposite differences cannot cancel. Closest matches have an average distance of at most 5 points. Higher/lower selections require the chosen gap against every active source. Split verdicts measure the distance between the highest and lowest external percentiles.

Spearman’s ρ measures how similarly we order the full comparison set: +1 is the same ordering, 0 indicates no rank correlation, and −1 is a reversed ordering. It is unavailable for a constant rating distribution.

IMDb reflects audience ratings, RT reflects the share of positive critic reviews, and Metacritic aggregates critic scores. Percentiles compare relative placement while the original scores remain visible in the details. These are saved score snapshots, including historical imports. Missing or doubtful values marked UNSURE are excluded.