Saturday, October 3, 2026 brings a lopsided number to the board: UAB is being asked to cover -31.5 against Samford, with no separate moneyline posted on that side. Any Samford vs UAB prediction built around a mega-favorite spread lives or dies on one question: does the favorite's own math actually support that kind of margin, or is the number just reflecting the gap in levels between the two programs?
The desk ran the numbers on this one and landed on PASS, an official-card no-action tier. We are not backing UAB -31.5, and we are not backing Samford either. Here is the reasoning, stated plainly, with the specific numbers that drove it.
The number that matters
The model's own projection has UAB winning this game straight up by 8.6 points. That is the baseline figure, with a 0.0-point adjustment applied on top of it, meaning no matchup overlay in this spot was established well enough to move the number off that baseline. An 8.6-point projected margin and a -31.5 market line are not in the same neighborhood: the spread is more than three and a half times larger than what the model itself thinks this game is worth. When a favorite's own projection sits that far under the number it is being asked to cover, the gap between the two is the headline, whatever is behind it.
Why the model still leans UAB
UAB's path to covering even its own smaller number runs through sustaining drives against a lower-level front, and the model notes UAB has turned in competitive results against Group of Five competition this season. The matchup is also a short, local trip for UAB, which the model treats as adding little in the way of travel burden. Ryder Burton's four starts give UAB some week-to-week continuity at the position, which the model counts as a small stabilizing factor rather than a reason to push the number higher.
The case against a blowout
On the other side, Samford's most recent scouting look, dated September 26, makes Gabarri Johnson the quarterback reference to use, not the older Crittendon reports the model treats as stale. Johnson went 22-of-26 against FCS opponent VMI, which speaks to his accuracy and offensive competence at that level. It does not by itself prove he can produce the same thing against FBS pressure, and the model is careful not to stretch one game further than it can support. Samford's history includes blowout losses to power-conference opponents, but the model will not borrow that result as proof that UAB's offensive line can dominate in the trenches the way a power-conference front did. It is a different level of competition, and the model keeps the two separate.
There is also a protection question hanging over UAB. Kirkland's return is unresolved, and the model does not credit Tate as restored interior depth. Reports that Spencer and Jackson are back at practice do not establish either full clearance or continued absence, so the model treats both as open questions rather than good news. A verified interview also flagged ongoing protection and passing-timing problems up front, along with unnamed, day-to-day injury designations elsewhere on offense. That is not enough, on its own, to say multiple starters are out, but it is enough to keep stalled drives a live issue. That is the mechanism behind the matchup: Johnson's efficiency keeps Samford on schedule on offense while UAB's protection and timing issues make it harder for UAB to pull away as fast as a 31.5-point number would require.
The verdict
Put it together and the picture is a favorite the model likes straight up by less than nine points, carrying real protection and timing questions up front, being asked to cover a number more than three and a half times larger than its own projected margin. That gap is enough for this one to land on PASS. We are not making this bet in either direction, not UAB -31.5 and not Samford as a live underdog story. The distance between the model's own number and the market's number is the story on this one, and it is one we are watching from the sideline on Saturday.
- Event · College Football Week 5 · Saturday, October 3, 2026
- 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.

