Slate Stats › NFL › Studies
Does a new head coach change who gets the ball?
Verdict — It changes WHO, not how muchYes, and almost entirely on one axis. How much a team throws barely notices a coaching change. WHO it throws to does not survive one at all — a tight end’s share of the targets carries from one season to the next at +0.55 under a stable staff and +0.07 after a change, which is indistinguishable from zero.
Every August someone points at last season’s target share to argue a tight end is due. It is a reasonable thing to look at — under a stable staff it is one of the more durable team traits in football. The question nobody asks first is whether the same people are still calling the plays.
| What carries to next season | Same coach | New coach | Gap |
|---|---|---|---|
| Pass volume (targets) | +0.45 | +0.30 | 1.2 se |
| TE share of targets | +0.55 | +0.07 | 3.8 se |
| WR share of targets | +0.47 | +0.08 | 2.9 se |
| RB share of targets | +0.48 | +0.23 | 2.0 se |
The split is the whole finding. Pass volume is mostly roster and game script — it carries at +0.45 under the same coach and +0.30 under a new one, a gap of 1.2 standard errors, which is nothing. A new staff inherits the same quarterback and the same schedule and throws about as often as the last one did.
Distribution is the staff’s own signature. Tight end share falls from +0.55 to +0.07, a gap of 3.8 standard errors. Wide receiver share falls from +0.47 to +0.08. Running back share holds up best at +0.23, which fits: checkdowns are forced by pressure and down-and-distance more than by preference.
The practical version is blunt. After a coaching change, a team’s historical position target distribution is close to worthless. Do not carry it forward, and be most sceptical about tight ends, where the collapse is largest and the position is thinnest. Ten teams changed head coach into 2026 — Arizona, Atlanta, Baltimore, Buffalo, Cleveland, Las Vegas, Miami, the New York Giants, Pittsburgh and Tennessee — and for those ten, last season’s usage chart is a description of a team that no longer exists.
How this was measured
Every team-season from 2016 through 2025 with more than 300 targets, paired with the following season for the same team, from official play-by-play. Head coach is taken from the schedule file for historical seasons and from a hand-confirmed staff table for the current one. Correlations are across team-seasons; the gap is a Fisher z test on the difference between the two correlations. n = 224 same-coach pairs and 62 new-coach pairs. A correlation from 62 pairs carries a standard error of about 0.13, which is why +0.07 is reported as indistinguishable from zero rather than as a small positive.