On/Off Splits: Measuring a Player's True Lineup Impact
What are on/off splits?
On/off splits measure how a team performs when a specific player is on the court versus off it. The core question: does the team score more, defend better, and win more often with this player playing than without them? This approach captures impact that individual statistics cannot — the teammate elevation, defensive organisation, and spacing effects that a player creates without ever touching the box score.
Reading on/off data
On/off data is typically presented as Net Rating — points scored per 100 possessions minus points allowed per 100 possessions. If a team has a Net Rating of +8 with Player X on the court and -3 with Player X off the court, that is an 11-point swing — Player X is clearly enormously impactful even if their raw statistics are modest.
This is how players like Draymond Green at Golden State became analytically celebrated despite relatively modest scoring averages. When Draymond was on the court, the Warriors' defensive and offensive organisation was dramatically better. When he was off it — often due to foul trouble — the team's performance declined significantly. On/off splits made this visible when traditional stats could not.
The problem: sample size and lineup quality
On/off splits are heavily influenced by who a player shares the court with. A reserve player who only plays with the starters will have great on-court numbers because their teammates are excellent. A starter who plays many minutes with weak bench lineups will have worse off-court numbers than their actual impact warrants.
True on/off analysis requires large samples (multiple thousands of minutes) and statistical adjustments for the quality of teammates and opponents faced. Single-season on/off splits can be dominated by noise and should be treated cautiously. Multi-year averages and adjusted plus/minus metrics (which control for teammates and opponents simultaneously) are more reliable.
Practical use for analysis
On/off splits are most useful for answering specific questions: Does this team's defence collapse without its defensive anchor? Does this team's offence improve dramatically when the ball-handler rests? When a player is injured, what on/off data exists that suggests how much impact their absence will have? Used as contextual evidence alongside other metrics, on/off splits are among the most informative tools available for understanding player value beyond the counting statistics.