Only counting games with fully reconciled play-by-play contribute assisted data. Individual appearances can be lower than team coverage.

Combined score = 49.5% playing style + 27% production + 13.5% career shape + 10% position similarity. Exact position groups score 100, overlapping hybrid positions score 85, adjacent roles (guard/forward or forward/center) score 50, and guard/center scores 0. These are fixed model choices. When either position is unknown, the position component is omitted and the other weights are renormalized. Each score is a descriptive index, not a probability or forecast. Production uses points, rebounds, assists, steals and blocks per game, with weights 40/20/20/10/10. Per-game values are used consistently across profiles; they are not pace or playing-time adjusted.

Style weights: scoring distribution (15%), three-point attempt share (11.25%), free-throw attempt rate (7.5%), assists per offensive action (11.25%), turnover share (7.5%), offensive rebound share (7.5%), effective field-goal percentage (7.5%), true shooting (7.5%), assisted-basket share (12.5%), assisted points as a share of scored plus assisted points (12.5%). Scoring distribution is the mix of points from two-point makes, three-point makes and free throws; shot selection uses attempts. Offensive actions = FGA + 0.44 FTA + turnovers + assists. Missing measures are omitted and shared weights are renormalized. At least 70% style coverage is required.

For each scalar metric, distance = min(1, absolute difference / the stated scale). Score = 100 × (1 − weighted mean distance). Scoring distribution uses half the sum of the three absolute share differences. The scales are visible in each comparison. They are fixed design choices, not fitted or validated thresholds.

Career shape sorts season box-production values from best to worst, then applies weights 1, 0.95, 0.90, and so on. Value = PPG + RPG + APG + SPG + BPG. Shape = max(0, 100 × (1 − 2 × weighted absolute difference / combined weighted values)). Unmatched seasons receive zero. This adapts Basketball Reference's shape method using a simple box-score proxy, not Win Shares.

Same-stage comparisons use the first N recorded seasons of each candidate and require at least N. The first recorded season may not be the first college season. The selected season determines which players appear in the player dropdown. The 2026-27 selection includes the full current roster. Until a player has published 2026-27 statistics, Selected season uses the player's latest recorded St. Edward's season as a clearly labeled preseason baseline. Players without recorded statistics do not receive invented scores. Career and same-stage comparisons use the selected player's recorded seasons through that year; later seasons are excluded from the target profile. Available-career opponents use all their recorded seasons and may cover incomplete careers. Selected-season comparisons search all available seasons for each opponent's best match. Players are excluded from their own matches. Minimum sample: 10 games and 30 FGA for each aggregate profile. Lower-volume profiles remain provisional.

Points assisted on are the exact two or three points scored on a basket credited to that passer. Assisted-basket share is the proportion of covered field goals credited with an assist. Each accepted game must match every St. Edward's player's field goals, three-pointers and assists in its box score. An assist is paired to the preceding made field goal in the same period, on the same team, with the same clock or a blank assist clock. Missing, ambiguous or mismatched play-by-play is excluded, never treated as zero. Assisted metrics require at least 10 validated appearances and 80% game coverage for both compared profiles. Coverage is limited to St. Edward's appearances; previous-school statistics are excluded. Older record-book leaders without complete profiles are excluded. Scores reflect statistical resemblance, not scouting judgments or predictions.

Basketball Reference method