The rushing team won 7 of 16 NFL Week 2 games. The passing-efficiency team won 13.

The ESPN scoreboard endpoint returned 33 completed 2026 NFL regular-season games through the Thursday night Week 3 opener: 16 games in Week 1, 16 in Week 2, and Thursday's Atlanta-at-Green Bay contest. Each game's box score was then pulled individually from the ESPN game-summary API by event ID on September 26, 2026. All counts below describe those 33 games as returned, not any team-season totals held elsewhere. Across those 33 games, the team that rushed for more total yards won 21 times (63.6%). The team that accumulated more net passing yards won 24 of 33 (72.7%). The team with the higher net yards per pass play — ESPN's yardsPerPass figure, which divides net passing yards by pass attempts plus sacks taken — won 25 of 33 (75.8%). Week 1 and Week 2 moved in opposite directions on the rushing metric. In the 16 Week 1 games the endpoint returned, the rushing-yards leader won 13 times (81.3%); the passing-efficiency leader won 11 times (68.8%). In the 16 Week 2 games returned, the rushing-yards leader won just 7 times (43.8%); the passing-efficiency leader won 13 times (81.3%). Nine Week 2 games were won by the team that rushed for fewer yards. In 7 of those 9 games, the winner also held the higher net yards per pass play. In the other 2 (Philadelphia over Tennessee, Green Bay over the Jets), the winning team had the lower passing-efficiency figure. Two of those 7 games show the split clearly. Cleveland rushed for 44 yards on 23 carries (1.9 yards per carry) and beat Tampa Bay, which ran for 137 yards on 25 carries (5.5 ypc), 23-19. Cleveland threw for 223 yards at 7.0 net yards per pass play; Tampa Bay threw for 176 at 4.8. Dallas rushed for 66 yards on 21 carries (3.1 ypc) and beat Washington 37-20. Washington ran for 182 yards on 32 carries (5.7 ypc) but threw for 207 yards at 5.9 net yards per pass play. Dallas threw for 266 at 8.1. Across all 7 of those games, the winners averaged 7.9 net yards per pass play and the rushing-heavy losers averaged 5.3. Thursday's Week 3 opener — Atlanta 35, Green Bay 14 — ran both metrics in the same direction. Atlanta rushed 41 times for 242 yards (5.9 ypc) and threw for 256 yards at 10.2 net yards per pass play. Green Bay attempted 53 passes for 311 net yards (5.8 net yards per pass play) and ran 9 times for 17 yards (1.9 ypc). Atlanta held both the rushing-yards edge and the passing-efficiency edge.

The following is analysis, not fact. The week-to-week swing in the rushing-yards winner-match rate — 81% in Week 1, 44% in Week 2 — is large enough to reflect noise ^sspeculative. The 16 games the endpoint returned in each of the first two weeks are a small sample, and early-season variability in scheme, weather, opponent, and health makes week-level rates unstable. This is a snapshot of the games the endpoint returned, not the full season. The 33-game pool can support a directional observation — passing efficiency tracked winning more consistently than rushing yards did — but it cannot establish whether efficiency causes outcomes or accompanies them ^sspeculative. One possible reason to distrust rushing yards as a predictor is game script ^sspeculative. A team trailing late may throw more and run less; a team leading late may run more to drain the clock ^sspeculative. If so, both behaviors can inflate the winner's rushing-yard advantage, or strengthen the observed association between rushing-yard leadership and winning — making that advantage partly a mirror of the score rather than an explanation for it ^sspeculative. Passing efficiency, measured as net yards per pass play, may also be affected by game state — trailing teams may see softer coverage and throw for higher yardage — but it may move less uniformly in that direction, since quarterbacks may face more pressure when behind, which could suppress efficiency ^sspeculative. The more consistent winner-tracking for passing efficiency is consistent with that account, though the box scores alone cannot separate the two mechanisms ^sspeculative. Net yards per pass play varied across the 7 Week 2 wins. San Francisco posted 13.0 against Miami (35-13); Minnesota won its 9-3 game with 5.5 to Chicago's 4.2. New Orleans beat Baltimore (6.3 vs. 6.0) with a gap of just 0.3 yards per pass play — a margin that would translate to meaningful yardage only over the Saints' full pass volume ^sspeculative. Thursday's Atlanta-Green Bay game splits the two readings. Atlanta's 242-yard rushing day alongside 10.2 net yards per pass play fits neither the desperation-artifact nor a clean causal story — it led on both metrics. Green Bay's 17 rushing yards on 9 carries against 53 pass attempts is the artifact pattern itself — a team that appears to have fallen behind early and largely stopped running ^sspeculative.

The most direct challenge to this framing is sample size. The 33 games the endpoint returned are 12% of the 272-game regular season. Week-to-week swings of 37 percentage points on the rushing-yards winner-match rate show how unstable single-week results are. A complete-season analysis — adjusted for game script, quarter-by-quarter scoring margins, and team quality — would be needed before any metric is credibly labeled "more reliable" in a causal sense. A second challenge: net yards per pass play as computed here does not adjust for coverage, defender alignment, or air-yards distribution. A team that compiles a high ypa figure by throwing short passes into soft zone coverage when leading by two scores may be posting a score driven by game state rather than by quarterback or offense quality ^sspeculative. Advanced metrics — EPA per dropback, CPOE, or Defense-Adjusted Value Over Average — attempt to partial out these effects. This piece does not have access to those metrics for 2026 and relies on raw efficiency figures only ^sspeculative. The conventional defense of rushing yards as a team-quality signal is that the ability to run the ball consistently reflects offensive-line strength and run-game execution that also underpins a team's overall ability to win. On that framing, the pattern is tracking team quality through a coarse proxy, not claiming rushing yards cause wins. Week 1's 81% rate is consistent with that view; on it, Week 2's 44% would read as the anomaly — though the 16 games the endpoint returned in each week cannot establish which rate is the norm ^sspeculative.