Player shooting

Each of a player’s games is labelled by how long he had rested — counted from the games he actually played, not his team’s schedule. A player returning from injury into a team back-to-back is rested; the team is not.

no rest      played again the very next day
3+ days rest  three or more days since his last appearance
rest effect   eFG% on 3+ days − eFG% on no rest

The two buckets deliberately do not cover everything. One and two days of rest are the overwhelming majority of games and sit between the two states being contrasted; including them would blur exactly the comparison the page exists to make.

Shooting is measured as effective field goal percentage, which credits a three-pointer at 1.5 times a two because it is worth 1.5 times as much. Raw field goal percentage would punish a high-volume three-point shooter for taking the more valuable shot.

eFG% = (FGM + 0.5 × 3PM) / FGA

A rest effect is a difference of two percentages, each from a limited number of attempts, so it carries real uncertainty. The page computes the standard error of that difference and uses it to decide what is worth bolding — a split inside the noise is drawn muted rather than left to look like a discovery.

standard error ≈ √( 2500 / noRestFGA + 2500 / restedFGA )

The 2500 is a deliberate simplification: it treats every shot as a coin flip at 50%, which is close enough to real eFG% to give an honest scale and avoids implying more precision than the data supports. The practical consequence is the important part — a player with 60 no-rest attempts needs an enormous split before it means anything, and most players never clear it in a single season.

The attempt floor is adjustable on the page. Lowering it surfaces more players and more extreme-looking effects, almost all of which are small samples rather than real differences — which is the lesson the control is there to teach.

• —Shot difficulty. A rested player taking harder shots against a set defence can shoot worse while playing better.
• —Why a player was rested. Games missed through injury look identical to games missed for load management, and a player returning from injury is not 'rested' in any useful sense.
• —Opponent quality and defensive scheme.
• —Minutes and fatigue within a game. This is about rest before tip-off, not late-game legs.
• —One season of any single player's split is mostly noise. The page shows season-by-season precisely so a reader can see how much a supposed effect moves year to year.