Skip to main content
ROTATION ABSENCE · POINTS OF MARGIN

Availability Cost

How do missing players relate to the final score margin? Compare historical associations, measured after games.

How this is calculated

Missing a top player

2.86 pointsLOWER FINAL MARGIN WITH A TOP PLAYER ABSENT · ASSOCIATION
Best player out2.86
Playing at home2.82
On a back-to-back1.76
Visiting altitude1.36
Off an overtime0.58

The best-player-out estimate is similar in size to the home-court estimate. Bars compare the size of associations, not their direction: home court is associated with a higher margin, while the other conditions are associated with a lower margin. All are measured in points across 35,458 games, controlling for both teams' prior records. These are regression associations, not isolated causal effects.

How often players are absent

17.1%OF GAMES HAVE ONE SIDE MISSING ITS BEST PLAYER
44.5%OF TEAM-GAMES ARE MISSING NOBODY FROM THE ROTATION
8.6PLAYERS IN A TYPICAL ROTATION

Top-player absences over time

Top-player absences are more than three times as common as in 1996-97. That is context for every figure above: the same absence was a rare event across most of this sample and is an ordinary one now.

1996-97 · 6.0%2025-26 · 19.5%

The climb is not steady. It dips in 2023-24, the season the league first required 65 games for awards eligibility, and returns above the old line the year after.

Schedule estimates after accounting for absences

Measured absences change the schedule estimates little. Adding absence value to the regression moves every schedule coefficient by under 8%. Unmeasured absences and changes in minutes can still matter.

Show coefficient comparison
SITUATIONSCHEDULE ONLYPTSABSENCE HELD FIXEDPTSSHIFT%
On a back-to-back1.7611.6426.7
Visiting altitude1.3581.2825.6
Off an overtime0.5770.5347.4
Schedule density0.2750.2653.8

Limits of the comparison

Availability is measured after the game. Every figure here is worked out after the fact, from who actually took the floor. Lineups are not settled until shortly before tip, so none of this forecasts a game.

Much of the margin remains unexplained. Final margins have a standard deviation of 13.6 points; the model's error after including team records, schedule, and absences is 12.4 points.

Absence is inferred from the rotation a team had actually been using, because a long-term injury can keep a player off the sheet entirely. Coverage starts in 1996-97, where the box scores this rests on begin.