We called Texas by two and Texas won by one, and nothing about the middle of that sentence was as calm as the ends. Ohio State led 23–3 with 15 minutes to play after three quarters in which Texas turned the ball over on its first two possessions, kicked a field goal from the 14 on its first trip inside the 20, and failed a fourth-and-1 at midfield. Then Arch Manning led three straight touchdown drives of 91, 72 and 73 yards, converting two fourth downs, Connor Hawkins missed from 45 with 4:16 left, and Hollywood Smothers scored from the 1 with 25 seconds on the clock. The 26–24 call was right on the winner and within a point on the margin. The market, at Texas by 1.5, was off by half a point. Our model said Ohio State by 2.4, the season's first split, and the market won it.
We said Texas would win by two. Texas won by one. The game was much wilder than that sounds. Ohio State led 23–3 with 15 minutes left. Texas had given the ball away on its first two drives, settled for a field goal the first time it got close, and failed on a fourth-and-1 at midfield. Then Texas scored touchdowns on three straight long drives, Ohio State's kicker missed from 45 yards with about four minutes left, and Texas scored the winning touchdown with 25 seconds to go. Our prediction had the right winner and was one point off on the margin. Our computer model had picked Ohio State. It was wrong, and the betting line was right.
17 falsifiable claims graded: the score call, three tells, thirteen matchup-board edges. Nine storylines graded on whether they affected the game or the read. Full grades live in the season ledger (data/2026/season_ledger.csv), with each pregame claim quoted verbatim.
Every wrong chip was about Texas throwing the ball. We called Coleman and Wingo against Sanchez and Mathews Even, the Texas deep game against Buckeye coverage Even, and Sayin's deep game against Texas coverage a slight Texas edge. The rebuilt Ohio State secondary held Manning to 5.3 yards an attempt and a longest completion of 21; Coleman and Wingo combined for 7 catches and 57 yards; and Sayin hit Jeremiah Smith for 43 yards twice. Three chips, one direction: we discounted a secondary that lost its best player and it did not need him for three quarters.
All three of our WRONG matchup calls were about Texas passing the ball. We said the Texas receivers and the Ohio State defensive backs were even, that Texas's long passes and Ohio State's coverage were even, and that Texas had a small edge covering Ohio State's long passes. None of that happened. Ohio State's defensive backs held Manning to 5.3 yards a throw with nothing longer than 21 yards, and Ohio State hit two 43-yard passes to Jeremiah Smith.
The Simmons tell printed the wrong quarter. We said a Simmons pressure or sack in the first quarter meant Ohio State would keep tight ends in to block. His first-quarter contribution was a tackle for loss on a run; the sack came with 1:54 left in the second. We grade the number we printed, and the number was WRONG, the same standard we applied to the Aggie sack tell this morning.
Our Colin Simmons prediction was WRONG. We said he would get to the quarterback in the first quarter. He did not. He tackled a runner behind the line in the first quarter and got his sack late in the second. We hold ourselves to what we printed.
The model picked the wrong team, and it was not close to close. Ohio State by 2.4 was a straight-up miss in a one-point game, which is exactly the situation a margin model should be judged on. The market's 1.5 was off by half a point. That is one split, decided the way the backtest said splits would be decided. It changes nothing about the rule and it goes in the December file as the first data point.
Our computer model picked Ohio State. That was WRONG. The betting line picked Texas by 1.5 and was off by half a point. This is the first game where the model and the line disagreed, and the line won. That is what our testing said would happen, so we keep using the line for our call.
Texas fumbled on its first snap, Raleek Brown losing the ball at his own 18, and Ohio State kicked a field goal. Manning's second drive ended with Jaylen McClain's interception at the Buckeye 12; Ohio State went 81 yards and Ja'Kobi Jackson scored from the 8. Texas's first real drive, 16 plays to the 14, produced Gianni Spetic's 37-yarder. Julian Sayin found Jeremiah Smith for 43 and then for the 19-yard touchdown with ten seconds left in the half, 17–3, after Texas had failed a fourth-and-1 at midfield. Hawkins's 48-yarder in the third made it 23–3 following the second 43-yard Smith catch. Then the game turned: a 91-yard drive that converted fourth-and-9 to Emaree Winston and ended with Ryan Wingo's 3-yard touchdown 41 seconds into the fourth; a 72-yard drive that converted fourth-and-6 to Emmett Mosley and ended with Smothers from the 1 at 7:14; a Hawkins miss from 45 at 4:16; a 73-yard drive and Smothers from the 1 again at 0:25. Graceson Littleton intercepted Sayin's last throw. Texas outgained Ohio State 239 to 132 after halftime.
| Texas | Ohio State | |
|---|---|---|
| Total yards | 336 | 372 |
| Passing / rushing | 196 / 140 | 278 / 94 |
| Third downs | 9 of 16 | 6 of 14 |
| Fourth downs | 2 of 3 | 0 of 1 |
| Turnovers | 2 | 1 |
| Points off the sideline's biggest fear | The red zone: a field goal from the 14 on the first trip, then three touchdowns from inside the 5 in the fourth quarter | |
Next week is UTSA. The Roadrunners visit DKR on Saturday, Sept 19, 7:00 PM CT. The scouting report and a frozen call publish Monday morning. Saturday told us the defense can strangle an elite offense for a half and the offense can win from behind on the ground; it also told us Texas gave a No. 1 team ten points on its first two snaps.
The first trip, a 16-play drive to the 14 in the second quarter, ended in a 37-yard field goal after a sack on second down. By the tell's own logic that meant not fixed, and Texas spent the next 30 minutes chasing the game, which is what the report said would happen. Then the fourth quarter: three trips inside the 5, three touchdowns, two of them Smothers from the 1. The tell read the first three quarters right and the ending wrong. PARTLY.
PARTLY right. The first time Texas got close, it kicked a field goal instead of scoring a touchdown, and we had said that would mean trouble. It did: Texas trailed for most of the game. But in the fourth quarter Texas got close three times and scored three touchdowns. So the problem was not fixed early, and then it was.
Forty-eight in the first quarter, on four catches, two short of the number. He finished with 8 for 164 and a touchdown, more than three times his previous two games against Texas combined, with two 43-yard catches. So the plan did fail and Ohio State did have its best player, exactly as the tell said it would, and Ohio State lost anyway, which the tell did not allow for. The threshold was two yards from misleading a reader at the quarter break, and the yards decided nothing. PARTLY.
PARTLY right. We said watch whether Smith has 50 yards after the first quarter. He had 48. He finished with 164 yards and a touchdown, far more than in his last two games against Texas. So Texas did not stop him, as our tell warned. But Ohio State lost anyway, and our tell assumed that if Smith got loose, Ohio State would win.
No Simmons pressure or sack in the first quarter; a tackle for loss on a run was his only first-quarter mark. He sacked Sayin with 1:54 left in the half, Lance Jackson had one two minutes earlier, and Cryer added two hurries after halftime as the Buckeye offense shrank to 132 second-half yards. The version Texas could live with arrived; the tell said it would arrive in the first quarter, and it did not. WRONG.
WRONG. We said Simmons would get to the quarterback in the first quarter. He did not. He got his sack late in the second quarter, and the Texas defense got better and better after halftime, holding Ohio State to 132 yards. The idea was right, and the timing we printed was wrong.
All thirteen pregame pairings, the edge we printed Thursday, and what the game said. Chips are quoted verbatim from the report.
| Pairing | Our call | What happened | Grade |
|---|---|---|---|
| The Quarterbacks: Manning vs. Sayin | Edge: Ohio State, slight | Sayin 17/32, 278, 1 TD, 1 INT (8.7/att); Manning 23/37, 196, 1 TD, 1 INT (5.3/att) and three fourth-quarter touchdown drives | Right |
| Coleman and Wingo vs. Sanchez and Mathews | Edge: Even | Coleman 3-35, Wingo 4-22-1; Mosley led Texas with 67; longest Texas completion 21 | Wrong |
| Texas OL vs. the Buckeye Interior (Run Blocking) | Edge: Ohio State | 140 yards on 41 (3.4), longest run 12, a failed fourth-and-1; the second half was better (103 yards) | Right |
| Smothers and Brown vs. the Buckeye Run Defense | Edge: Ohio State | Smothers 21-74-2, Brown 10-54 and a lost fumble; 4.1 a carry between them, two touchdowns from the 1 | Partly |
| Texas Pass Protection vs. the Buckeye Pass Rush | Edge: Ohio State, slight | 2 sacks (Pettijohn, K. Jackson) and 3 hurries on 39 dropbacks, all before the fourth quarter | Right |
| Texas Deep Game vs. Buckeye Coverage | Edge: Even | No Texas completion longer than 21; 5.3 yards an attempt | Wrong |
| Jeremiah Smith vs. the Texas Secondary | Edge: Ohio State | 8-164-1 with two 43-yard catches; 48 yards in the first quarter alone | Right |
| Colin Simmons vs. the Buckeye Tackles | Edge: Texas, slight | 1 sack (Q2), 1 TFL; Texas 2 sacks and 4 hurries as a team; Sayin still 8.7 an attempt | Partly |
| Bo Jackson vs. the Texas Front | Edge: Texas | 19-74 (3.9), long 11; Ohio State 3.0 a carry overall and stuffed on fourth-and-2 at the Texas 12 | Right |
| Buckeye OL vs. the Texas Front (Run Blocking) | Edge: Texas, slight | 94 yards on 31 (3.0); 46 second-half rushing yards | Right |
| Sayin's Deep Game vs. Texas Coverage | Edge: Texas, slight | Two 43-yard completions to Smith, 8.7 an attempt, 278 yards | Wrong |
| Special Teams & Field Position | Edge: Texas | Texas started at its own 21 on average, Ohio State at its 36, mostly on Texas's own turnovers; Hawkins 3/4 with the decisive miss, Spetic 1/1 | Partly |
| Coaching & Luck | Edge: Even | A one-point game: turnovers 2-1 Ohio State, a Buckeye fourth-and-2 stuffed at the 12, two Texas fourth downs converted, one missed 45-yarder | Right |
Reading the grades honestly: seven of thirteen edges were right, and the three wrong ones all sat on the same assumption, that a secondary missing an eleventh overall pick would leak. It did not leak for three quarters, and when Texas finally moved the ball it was on the ground and underneath. The board's best call was the one that made the line: Ohio State's rebuilt offensive line against the Texas front, where two slight Texas edges produced 3.0 a carry and a stuffed fourth-and-2.
Our model said Ohio State by 2.4 (28.3–25.9); the market said Texas by 1.5. Actual: Texas by 1. Market error 0.5, model error 3.4, and the winner went to the market. This was the season's first split, printed on Thursday for exactly this purpose. The season line is now model 2, market 2 on margin and market 1-0 when they disagree. The published call stays market-anchored.
For the record, the base rate we printed said favorites of 1 to 3 points went 111-115 and home favorites in that band 60-66: a coin flip. The coin landed on the home side by one point. Raw rows: predictions.csv.
The film behind the grades. Every link below was verified live before publish; anything we could not verify was dropped rather than guessed at.
No. 4 Texas grinds No. 1 Ohio State to a halt for epic comeback (ESPN). The three fourth-quarter drives, the Hawkins miss, and the Smothers plunge.
Football earns historic 24-23 comeback win over No. 1 Ohio State (Texas Athletics). The 105,044 attendance, the comeback records, Sarkisian's "heart, the desire and the fight."
Texas 24, Ohio State 23 (AP via ESPN). Day's "We had them on the ropes and didn't finish," Manning's "I knew it was our time."
YouTube highlight cuts: none passed link verification by press time; when ESPN or Texas Athletics posts one, the Monday edition carries it.
How the game is landing across the sanctioned doors, as of Sunday morning. These are conversations, not our claims; our claims are graded above.