A ranking is a question, not a verdict.

The best player board does not tell a coach whom to recruit. It makes the evidence behind a shortlist easy to inspect, compare and challenge.

The NCAA rankings boardoffers 799,620 source-linked player rows across counting stats, rates, possession context and lineup impact. Every row keeps its NCAA player and team IDs, season, games and minutes visible. Those fields establish what the release recorded; they do not establish a current roster spot, eligibility or a future role.

Choose the question before the metric

Use points per game or points per 40 when the question is scoring volume, then check minutes and games. Use true shooting or effective field-goal percentage to study shot efficiency, and read the matching attempt denominator beside the rate. Assist-to-turnover ratio, turnover rate and team possession share describe ball security and involvement; they are not interchangeable definitions of creation. Rebounding and stocks per 40 can surface a role, but they still need lineup and matchup context.

Use the two composite views as screens

The balanced production index standardizes eight available components— points, rebounds, assists, steals, blocks, true shooting, effective field goal percentage and points per 40—within the filtered board and averages the components that meet their denominators. At least four components are required. It is useful for finding broadly productive players, while its component audit shows what is missing.

The impact + production index requires an exact-ID net RAPM record and points per 40. RAPM must include at least 500 offensive and 500 defensive possessions. That makes the screen narrower and more useful for a lineup question, but a missing RAPM row is missing evidence, not zero value. Open the impact board to inspect the possession sample before drawing a conclusion.

Turn a rank into a recruiting workflow

Start with a role: secondary creator, spacing wing, defensive rebounder or point-of-attack defender. Filter the ranking to a meaningful sample, open the source player card, and inspect the game-level record. Then search the dated recruiting file and fit board. A school announcement can confirm what was published and when; it cannot by itself confirm availability, eligibility or the minutes a player will receive.

Keep the denominator in the report

A shortlist should name the season, source release, sample thresholds, metric and missing fields alongside each player. Compare similar roles and competition where possible. If a player ranks highly on a rate but has a small attempt or possession sample, write that limitation into the scouting note. The model notebook keeps these descriptive rankings separate from the 2026–27 matchup forecast, so a leaderboard never silently becomes a prediction input.

Data attribution: SportsDataverse, CC BY 4.0. See the source receipts and model notebook.