We read the 80 Week 2 college games ESPN listed for September 12: yards led 14 of 31 close games — turnovers led 18 of 22

The 80 college football games ESPN listed as Final for September 12, 2026 — the main Saturday slate of Week 2 — split into 31 results decided by 14 points or fewer, 41 decided by more than 21 points, and 8 in between. In the 31 close games, the team with more total yards won 14 and lost 17: a 45% correct rate. In the 41 blowouts, the yardage leader won 39 of 41 — 95% correct. Turnover margin held differently across both groups. Of the 22 close games where the two teams turned the ball over a different number of times, the team with fewer turnovers won 18 — 82%. Nine close games ended with both teams posting the same turnover total. Among the 32 blowouts where turnover counts differed, the team with fewer turnovers won 26 — 81%: essentially the same rate as in tight games. One of the starkest inversions came in the Carrier Dome, where the California Golden Bears beat the Syracuse Orange 21-18 with 249 total yards against Syracuse's 437 — a 188-yard gap — while turning the ball over zero times to Syracuse's 3. Ohio University beat Jacksonville State 29-27 with 262 yards against Jacksonville State's 447 (a 185-yard gap), with Ohio posting 0 turnovers against Jacksonville State's 1. The one exception: Texas 24, Ohio State 23. Texas won with fewer yards (336 to 372) and more turnovers (2 to Ohio State's 1) — the only close result that went against both indicators. Texas converted 9 of 16 third-down attempts (56%) against Ohio State's 6 of 14 (43%), and went 2-for-3 on fourth down while Ohio State went 0-for-1. Across all 80 games, teams averaged 1.36 turnovers per side.

The following is analysis, not fact. In blowouts, yards and outcomes moved together in this slate. When one team outproduces the other across the board, more yards and a large winning margin may be close to the same thing ^mmodeled. The 95% rate is consistent with one-sided domination rather than with yards causing the wins ^mmodeled. In close games, both offenses generate meaningful production. The gap in total yards between the two teams is typically much smaller. At that scale a team that produces more yards may also be running more plays, which may create more opportunities to turn it over ^mmodeled. The California-Syracuse and Ohio-Jacksonville State games are consistent with that reading: the team that generated more yards also turned the ball over more, and those yards appeared on the box score but not on the scoreboard ^mmodeled. A 437-yard offense that loses three possessions to turnovers does not score 437 yards of value ^mmodeled. The near-identical turnover rate in tight games and blowouts — 82% versus 81% — is consistent with turnovers tracking something about scoring efficiency regardless of how competitive the game is, while total yards may shift meaning between contexts ^mmodeled. In a rout, yardage and the win may tell one story ^mmodeled. In a close game, a yardage edge may reflect volume — more plays run — as much as dominance ^mmodeled. Texas-Ohio State is the exception to both reads. Texas won with a yards deficit and a turnover deficit. In a one-point game, what appears to separate the teams is down-by-down execution ^mmodeled: Texas converted on 9 of 16 third downs and twice on fourth down, which may have kept its drives alive when Ohio State was going three-and-out ^mmodeled. The box-score summary points one way; the conversion math points the other ^mmodeled.

A single Saturday of 80 games is a limited sample. Over full seasons, total yards is generally reported to correlate strongly with winning percentage across hundreds of FBS and FCS games ^mmodeled; one week's close-game inversion is a data point, not a structural rule. The turnover win rate of 82% in close games comes from only 22 games where turnover counts differed; 4 failures in 22 tests means small-sample variance could move the rate several points in either direction. Broadcast emphasis on total yards may reflect a metric that tracks team quality across a season ^mmodeled — the specific breakdown in games decided by two scores or fewer does not overturn that.