Read impact with the lineup context intact.
A player’s box score records what happened while they were on the floor. Adjusted plus-minus asks a harder question: how did the team’s scoring balance change after accounting for the teammates and opponents in those lineups?
The impact board publishes NCAA-derived regularized adjusted plus-minus (RAPM) from the attributed SportsDataverse release. ORAPM describes the offensive component, DRAPM the defensive component, and net RAPM is their sum. The publisher fits a ridge model across Division I stints, which helps keep a player with a small or highly unusual lineup sample from dominating the fit.
Start with the possession denominator
A large positive estimate with 2,000 possessions has a different level of evidence than the same estimate with 250. The board’s qualified view requires at least 500 offensive and 500 defensive possessions. That threshold is a browsing rule, not a guarantee that the estimate is stable. Use the possession columns to keep the sample visible when you sort by ORAPM, DRAPM or net RAPM.
Keep offense and defense separate
Net RAPM is useful for a first pass, but it can hide a player’s role. A strong ORAPM with a neutral DRAPM suggests a different film question than a defensive specialist with the reverse profile. Sort each component, then compare the result with the box-score archive for minutes, shooting, rebounding and turnovers. These measures answer related questions and should not be added into a new composite without a separate validated model.
Do not turn a historical estimate into a roster claim
RAPM describes recorded stints in a source season. It does not establish a player’s current team, health, eligibility or expected role next season. NCAA source IDs remain in their own identity namespace; they are not joined to ESPN identities by name alone. For 2026–27 preparation, put the estimate beside the dated recruiting evidence and current roster observations, then confirm availability with the school.
Use it to choose film, not to skip film
A useful workflow is to identify a player whose component estimate changes the matchup, inspect the possession sample, and then write a question for film. Does the offensive value come from creation, shot selection or finishing? Does the defensive value survive different matchups? The game-plan workbench turns those questions into matchup preparation while keeping the forecast model’s roster limitations explicit.
Data attribution: SportsDataverse, CC BY 4.0. See the source receipts and model notebook.