How we got there: Whose sacks are they?

We asked: when a quarterback gets sacked, whose number is that, his or the line's?

Football runs the experiment for us: quarterbacks change teams. The player_season_pairs view carries a changed_team flag for exactly this split. Sample: 150+ dropbacks in both seasons, 1999 to 2025.

Movers first. Sack rate before and after a team change. If the line owned the number, a new line should scramble it.

The baseline: stayers. Same test, quarterbacks who kept their team.

The control. Run the identical mover test on the rest of the stat line. Yards per attempt (below), then completion percentage over expected. In our published run, yards per attempt kept 41 percent of its stayer stability (r 0.18), CPOE kept 63 percent (0.32, charted from 2006), and sack rate kept 70 percent (0.35 against a 0.51 stayer baseline). The stats that travel are the ones the man carries: his clock and his accuracy.

The reverse experiment (teams that keep the roster but change the primary quarterback retain only r 0.11 of their sack rate, against 0.53 keeping him) aggregates by team and dropback leader, which is a scripted analysis; spec and results in the article backtest block.