North Texas vs Tulsa prediction season is here, and the model's math says this one shouldn't be close. The desk's projection has North Texas winning outright, with a margin that would make most bettors reach for their wallets. But the number on the board doesn't match the number in the model's head, and that gap is the whole story heading into Friday, October 2, 2026.
This is a PASS on the official card. No official action is going out on North Texas vs Tulsa, even with the model leaning green. Here's why the projection likes one side, and why the desk isn't backing it with money.
Why the model is on North Texas
The model's number is North Texas -11.5, meaning it projects North Texas to win by roughly two touchdowns. The path there runs through Jackson's passing production and White's rushing threat, the two-man engine that's carried the offense this season. The desk weighs the 44-6 win over UNLV, an FBS opponent, more heavily than the 63-14 rout of an FCS team, since beating a team at your own level tells you more than beating one that doesn't belong on the schedule.
On the moneyline, North Texas sits at -110, which is close to a coin flip in pricing terms for a side the model has winning by double digits. That gap between the model's projected margin and the market's moneyline is part of what makes this matchup interesting, even with no bet attached to it.
The case against
North Texas also allowed 49 points to Texas State, and the desk isn't pretending that result doesn't exist. A defense that gives up 49 in a single game has a real floor problem, and Tulsa's own win path runs directly at that weakness. The model credits Tulsa's 24-10 home win over Oklahoma State as its best evidence, a game where the Tulsa defense controlled tempo and limited a Big 12 offense. The six-point win over Sam Houston offers less separation, and the loss to Arkansas doesn't tell the model much either way, since an SEC loss doesn't establish a trench disadvantage against a Group of Five opponent like North Texas.
If Tulsa can repeat anything close to that Oklahoma State performance, the model's risk notes point to a path where the Tulsa defense limits Jackson's possessions and exploits the same North Texas defense that cracked against Texas State.
Why this is a pass
There's a quarterback wrinkle worth tracking even though it doesn't move the number. A report on September 30 restores a quarterback option for North Texas, but that report doesn't establish full recovery from a knee issue, so the model isn't treating it as a clean bill of health. Jackson's continued participation since then supports continuity in the offense without proving the earlier leg limitation is fully resolved. On the roster ledger, North Texas is without Jenkins and has dismissed Becks, while Tulsa gets no assumed recovery credit for Green or Thompson, leaving questions at linebacker and in the secondary for both sides.
None of that is enough for the model to tack points onto its own projection. It's holding at -11.5 rather than double-counting absences that are already baked into recent form. That's the core of why this stays a pass: the model has a lean, but not one it's willing to defend at today's price with real conviction. We are not backing North Texas vs Tulsa on the card today.
The verdict
The model's projection favors North Texas straight up, built on Jackson-and-White offensive production against a Tulsa defense that hasn't faced an attack quite like this one. But a North Texas defense that allowed 49 points to Texas State is a real crack in the armor, and Tulsa's Oklahoma State performance shows exactly how a defense can exploit it. That tension is why this one stays off the official card. The desk is watching how North Texas's defense holds up and how much the restored quarterback option actually changes the offense, but for now, this is a pass, not a play.
- Event · College Football Week 5 · Friday, October 2, 2026
- Market · North Texas -110 — recorded when the pick was posted
- Model · Heismann, Botcappers' AI College Football model — how it is built and what it does not know
- Record · every pick is timestamped before the event and graded in public, win or lose
- Statistics · the model's own match/game database, current to the posting time above. Figures a reader cannot open here are internal to that database, not published third-party data.

