CFB Sportsbook Consensus Grading Recap — Friday, September 25, 2026

Five games. 125 logged picks. 60 experts and models graded against the confirmed finals. Here is who cashed on Friday night.

SportsGrid Model had the best day on the board at 5-2, +3.66 units, built on the single biggest winner of the night — UAB on the moneyline at +202. Anthony Rome of TheSpread swept Army–Temple three ways for 3-0, +2.44 units: the side, the total and the moneyline, all correct. Philip Wood of Sportsbook Review went 3-1 (+1.71u) and Al MacMillan of Winners and Whiners went 3-1 (+1.65u), which moves MacMillan to 17-11 (+4.09u) on everything we have graded him on in CFB this season. Kyle Kargel of Picks and Parlays was 4-2 (+1.62u) across four games. Three College Football News cappers — Andre Cherry, Doug Millen and Jackson Langendorf — went a perfect 2-0 (+1.90u) by taking Clemson in a spot where most of the room did not.

The consensus card went 1-4 (-3.09 units). No STRONG play was posted on Friday; all five staked plays were LEANs, and the only winner was the Under in Clemson–California. The Temple FADE-WATCH also lost, carrying 0 units as watches always do. The full numbers are in the tables below.

What actually happened

Clemson 24, California 10. The consensus case for Cal ran through quarterback play, and it was half right: Clemson’s freshman Tait Reynolds, making his first career road start, went 13-of-21 for 108 yards with an interception. The problem was that Cal was worse. The Golden Bears did not score an offensive touchdown all night — their 7 came on a scoop-and-score and the rest on a field goal — while cornerback Elliot Washington II picked off two passes. Gideon Davidson ran for 159 yards and two scores on 22 carries, including a 75-yarder, and Bryant Wesco Jr. returned a punt 65 yards for a touchdown, Clemson’s first punt-return score since 2018. One fact explains both Clemson–Cal results on the card: neither offense could move the ball, which is exactly why the Under 50.5 cashed by 16.5 points — and it is the same reason the Cal side and the Cal moneyline never had a pulse.

Indiana 29, Northwestern 23. Indiana was up 12-0 at the half and never trailed, but the number was 20.5 and this was a backdoor of the ordinary kind rather than a garbage-time fluke. Aidan Chiles was 17-of-21 for 237 yards after the break with two touchdown passes to Griffin Wilde and a 7-yard rushing score; Luke Dehnicke caught seven for 140. Jackson Kleather’s 42-yard field goal inside the final minute made it a six-point game, and Indiana recovered the onside kick to end it. Indiana is now 0-4 against the spread this season while winning every week — the exact pattern the card flagged before kickoff.

UAB 24, Navy 20. Worth being precise here, because a voided premise and a beaten handicap are not the same thing. Navy’s starter Braxton Woodson was already out with the ankle injury that moved this number from about -11.5 to -6.5, and the experts on Navy argued openly that the market had overcorrected for a system offense. On the evidence, they were right about the offense: backup Jackson Gutierrez ran for 134 yards and a touchdown in his debut and Navy led 20-17 in the fourth. What beat the number was ball security. Gutierrez was sacked and fumbled at the Navy 48; UAB recovered and Jamarion McReynolds broke a 48-yard touchdown run on the next snap. Michael Barrow fumbled again at the UAB 44 to end the answer. Two fourth-quarter fumbles, not the quarterback situation, decided this one.

Army 21, Temple 17. The half-point mattered exactly as the FADE-WATCH note said it would, just in the wrong direction. Temple led 17-14 and outgained Army 459-315, but Cale Hellums scored his third touchdown of the night from three yards out with 1:41 left. An Army field goal there and Temple +3.5 cashes; a touchdown made it four. Temple’s last chance died on a fourth-and-3 incompletion to Jayce Freeman at the Army 23 with 29 seconds on the clock. Sam Brown had 96 rushing and 78 receiving yards for the Owls; Hellums carried 22 times for 104 and the three scores.

Friday by tier

Tier	Record	Hit rate	Units
STRONG	—	—	No STRONG plays posted
LEAN	1-4-0	20.0%	-3.09u
FADE-WATCH	0-1-0	0.0%	0.00u (watch only)
Card (STRONG + LEAN)	1-4-0	20.0%	-3.09u

By market, the LEANs were 0-3 (-3.00u) on spreads, 0-1 (-1.00u) on the moneyline and 1-0 (+0.91u) on totals. Every play is graded at the exact line and price we posted it at, win or lose.

Every play, graded

	Game	Consensus play	Tier	Final	Units
✅	Clemson @ California	Under 50.5 (-110)	LEAN	Clemson 24-10 California (34)	+0.91
❌	Clemson @ California	California +1.5 (-115)	LEAN	Clemson 24-10 California	-1.00
❌	Clemson @ California	California ML (-105)	LEAN	Clemson 24-10 California	-1.00
❌	Northwestern @ Indiana	Indiana -20.5 (-115)	LEAN	Indiana 29-23 Northwestern	-1.00
❌	Navy @ UAB	Navy -6.5 (-124)	LEAN	UAB 24-20 Navy	-1.00
❌	Army @ Temple	Temple +3.5 (-120)	FADE-WATCH	Army 21-17 Temple	0.00

Best call. SportsGrid Model on UAB’s moneyline at +202. The model had the Blazers as a live home dog against a service academy down its starting quarterback, and UAB won it outright on McReynolds’ 48-yard fourth-quarter touchdown after a Navy fumble at midfield — a +2.02-unit return on the single best price anyone took all night.

Worst beat. Navy -6.5. Navy led 20-17 in the fourth quarter with a backup quarterback running for 134 yards, which is the outcome the side needed, and then fumbled twice inside the last twelve minutes — once at its own 48, once at the UAB 44 — and lost outright. The handicap was not the thing that broke; the ball security was.

Expert scoreboard — Friday, September 25

60 distinct experts and models logged 125 picks across the five games. The field went 53-72 (42.4%). The consensus card went 1-4 (20.0%) — the field beat the card by a wide margin on the day, and the reason is almost entirely Clemson–California, where the card sat on the side most of the room was on and the room was wrong.

Top 5 — Friday	Day	Units	Season with us
SportsGrid Model	5-2-0	+3.66	8-10 · 44.4% · weight 0.94
TheSpread – Anthony Rome	3-0-0	+2.44	4-2 · 66.7% · provisional
College Football News – Andre Cherry	2-0-0	+1.90	3-0 · 100% · provisional
Sportsbook Review – Philip Wood	3-1-0	+1.71	6-6 · 50.0% · provisional
Winners and Whiners – Al MacMillan	3-1-0	+1.65	17-11 · 60.7% · weight 1.09

Bottom 5 — Friday	Day	Units	Season with us
Covers – Geoff Clark	0-4-0	-4.00	3-6 · 33.3% · provisional
College Football News – Pete Fiutak	0-4-0	-4.00	1-5 · 16.7% · provisional
Bleacher Nation	1-4-0	-3.09	9-7 · 56.3% · weight 1.03
TheRX – Ryan Callahan	2-4-0	-2.18	2-4 · 33.3% · provisional
Covers Computer Model	1-3-0	-2.09	1-4 · 20.0% · provisional

Season leaders — CFB ledger	Record	Units	Weight
CBS Sports – Chris Hummer	16-6-3 (72.7%)	+8.56	1.18
CBS Sports – Brandon Marcello	13-7-3 (65.0%)	+4.83	1.11
SportsLine Model	10-5 (66.7%)	+4.12	1.10
Winners and Whiners – Al MacMillan	17-11 (60.7%)	+4.09	1.09
Bleacher Nation	9-7 (56.3%)	-1.08	1.03

How this changes the card

Friday moved the weights, and it moved them in a way worth reading. Chris Hummer took Northwestern +21 and won; he is now 16-6-3 on everything we have graded him on in CFB this season, 72.7% on decided picks, and his weight ticks up from 1.17 to 1.18 — the highest any capper carries in our CFB consensus. Brandon Marcello was on the same side and moves from 1.09 to 1.11 at 13-7-3. Al MacMillan had the biggest single-day move among rated cappers, 3-1 on the night lifting him from 1.06 to 1.09 at 17-11.

Two more crossed the 15-decided-pick threshold on Friday and are now rated rather than provisional: SportsLine Model enters at 1.10 (10-5, +4.12u) and Bleacher Nation at 1.03 (9-7). Going the other way, SportsGrid Model also qualified — and despite owning the best single day on this board, its season sample is 8-10, so it enters the ledger at 0.94, below neutral. A great night is not a rating; 18 graded picks is.

The practical consequence for Saturday’s card is specific. The Northwestern +21 result is exactly the split the card flagged in its own notes before kickoff — the two cappers we have graded best in CFB were on the opposite side of a posted LEAN, and they were the ones who were right. With both weights rising and several Indiana-side cappers drifting below 1.00 — Chip Patterson and Brad Crawford to 0.94, Richard Johnson and David Cobb to 0.90, Shehan Jeyarajah to 0.85 — a similar split on a big spread will now tier lower, or not make the card at all.

Season to date — CFB consensus

Tier	Spread	Moneyline	Total	All	Units
STRONG	3-5	1-2	3-3	7-10	-3.90u
LEAN	5-8	0-3	3-3	8-14	-6.74u
Card (STRONG + LEAN)	8-13	1-5	6-6	15-24	-10.64u
FADE-WATCH (watch only)	1-2	—	—	1-2	0.00u

Across the last seven graded CFB dates, STRONG plays are 4-5 (-1.59u) and the card overall is 7-11 (-4.86u). Seven CFB dates have been graded this season. Those are the real numbers, posted whether they flatter the consensus or not — that is the entire point of grading it.

Grading notes

All five Friday games are confirmed final with no overtime, no postponements and no pushes. Every card play and every expert pick is graded at the exact line and price stored when it was logged; where a capper published a number that later moved, the published number is the one graded, and rows with no posted price are graded at -110. Units are flat 1-unit stakes. FADE-WATCH plays are graded and reported but carry 0 units and are excluded from the card record. Full receipts live in the graded sheets behind this recap.

Sources: ESPN, CBS Sports, NCAA.com and the participating schools’ official athletics sites for confirmed finals and game detail.

Betting involves risk and no record guarantees future results. Bet only what you can afford to lose. If gambling stops being fun, call 1-800-GAMBLER.