We read all 80 games ESPN's college-football scoreboard listed Saturday: halftime leads held 79% overall — but just 47% in close games

The team leading at halftime won 62 of 78 college football games on September 12, 2026 — a 79.5% conversion rate across the 80 completed games ESPN's college-football scoreboard listed that day, excluding the 2 games tied at the break. That number does two very different jobs. Among the 43 blowouts in this 80-game set — games decided by 21 or more points — the halftime leader won 41 of 43 (95.3%). Among the 20 games that finished within seven points, the 19 with a halftime leader saw that leader win 9 — 47.4% (the 20th was tied at the break). These one-score games are grouped by final margin, not by how close they were at halftime; a blowout that tightened late or a close game that broke open would be classified the same way. The 17 games decided by 8 to 20 points fell in between: 12 of the 16 with a halftime leader won — 75.0%. (One of the 17 was tied at the break and excluded from the denominator.) The second quarter is where scoring concentrated in this set. Across the 80 games ESPN's scoreboard listed, Q2 averaged 19.1 points per game — 35.4% of the 4,310 points scored in regulation (the set's full total was 4,346, including 36 overtime points across two games). Q1 and Q3 each averaged 11.4 points per game; Q4 averaged 12.1. Q2 outpaced each of the other three quarters individually by at least 58%. Ohio State at Texas was the day's most extreme exception. Ohio State led 20-3 at halftime. Through three quarters it led 23-3. Texas scored 21 unanswered points in the fourth quarter — the Longhorns' entire second-half scoring in one 15-minute span — and won 24-23. Ohio State scored nothing after the third quarter. The 17-point halftime deficit was the largest overturned in the 80-game set. Of the 16 comebacks in the 80 games ESPN's scoreboard listed, 8 erased halftime deficits of 3 points or fewer — a single score or less at the break. Five came from deficits of 7 or more: Texas at 17, Central Michigan at 10 (trailing Colgate), UTSA at 10 (trailing Texas State), South Florida at 7 (trailing Army), and UAB at 7 (trailing UL Monroe). Fourteen of the 16 comeback teams won Q4; 10 won Q3. Home teams went 63-17 across the 80 games in this set (78.8%).

The following is analysis, not fact. The 79.5% halftime-to-winner rate looks like a rule. In blowouts it appears close to one — 95.3% is consistent with games that were largely decided before the half, where the lopsided halftime score may be as much a trailing indicator of team quality as it is a predictor ^mmodeled. The gap between those teams' first and second halves is probably too large for halftime adjustments alone to explain; it more likely reflects a talent differential already visible before the break ^sspeculative. In the games that finished within one score, the 47.4% rate — 9 of 19 — is consistent with no predictive signal at all ^mmodeled. This group is defined by final margin, not by how close the game was at halftime, so it mixes games tight throughout with games that only closed or broke open late. The halftime leader and the halftime trailer appear to be on roughly equal footing in these games, with the board at the break not identifying which. ^sspeculative What appears to distinguish them is the fourth quarter: 14 of the 16 comeback teams won Q4, including 9 of the 10 one-score reversals (Tulsa the exception; its lead came in Q3). In these one-score finishes the result looks more tied to Q4 play than to the halftime position. ^sspeculative The Q2 scoring concentration (35.4%, nearly 68% more per game than Q1 or Q3) is consistent with teams pressing to score before the break and with the final possession of the half receiving explicit structure in both coordinators' game plans ^sspeculative. Whatever the Q2 concentration reflects, it does not appear to carry into the second half: Q3 averaged exactly what Q1 did (11.4 points per game), and the second half's scoring distribution is roughly flat ^mmodeled. The practical upshot: the halftime score appears to have anticipated the outcome in more than half of the games in this 80-game set — the blowouts, which made up 53.8% of it ^mmodeled. In the games that finished within one score, it did not — though that group is defined by final margin, not by how contested the game was at the break.

The 80 games are the full set ESPN's scoreboard listed for September 12, not a projection across the season. The 9-of-19 one-score halftime rate is a thin count; a full season's worth of evenly matched games would likely move that number and could shift it materially in either direction. This 80-game set also skews toward power-conference home games and scheduled mismatches — a composition consistent with the high blowout count (43 of 80) and a smaller pool of genuinely competitive games ^mmodeled. The Ohio State at Texas result — the set's headline outlier — is a single data point against the general rule, not evidence the rule doesn't hold; teams trailing by 17 at halftime rarely win ^sspeculative, and the Longhorns' Q4 required Ohio State to go scoreless in the fourth quarter, which is itself unusual. The home-field advantage (63-17 across the set) also entangles with the halftime conversion rate in ways the top-line figure does not resolve.