Week 4 Spreads the Market Got Wrong

How the board ranked

Plenty of games this week had real model edges - Wisconsin at Penn State, Georgia Tech at Stanford, South Florida at Bowling Green, Texas A&M at LSU, New Mexico at New Mexico State. The three core picks below all cluster in spreads of 7.5 points or less, where the dual-model disagreement with the market matters.

The board

1. Tulsa Golden Hurricane +7.5 vs. Arkansas Razorbacks. Sat 9/26, 8:00 PM ET

Best line: 7.5 at FanDuel (-122) · Confidence Score: 8.0 / 10

Arkansas is 0-3 ATS in its last three, averaging a 15.7-point miss against the number, and new coach Ryan Silverfield (in for neck brace legend Bobby Petrino) is still figuring out this roster. After coaching in a fake conference for 5 years, Silverfield says he's going to turn the team around, but we've spent too many seasons watching the Memphis Tigers tackle with the form of undersized middle school children. QB KJ Jackson has thrown for 618 yards with 4 TDs and 2 INTs, and the defense ranks just 105th nationally.

Tulsa counters with a defense ranked 75th and a QB in Baylor Hayes who has been modest (353 yards, 1 TD, 1 INT), which is probably because he looks 14 years old. Tulsa came out strong in Week 1, beating Oklahoma State outright, followed by a trash performance against Sam Houston, then by a whooping of East Texas A&M (not a real school) where Hayes did not play because he was grounded for getting a C on his homework. They are 2-1 ATS over their last three with a plus-7.3 average cover.

Thesis. Betting on Arkansas right now is like catching a falling knife. They come off two tough opponents in Georgia and Utah, but only put up 31 on North Alabama, who somehow managed to score 14 points on zero offensive production. Baylor Hayes is doubtful for this matchup, which we see as a positive, given Dexter Williams may have that dog. It's very possible that Traquan Williams runs train on Arkansas' "defense."

The numbers. SP+ has this at Arkansas -3.3, and FPI has it at Arkansas -1.8, against a market line of -7. That is a 3.7-point dual edge, with a projected margin of 2.5.

What breaks this. Arkansas's defense has time to turn a corner under a first-year staff, and they are an SEC team that will be facing much stiffer competition this season.

2. Troy Trojans +3.5 vs. Utah State Aggies. Sat 9/26, 7:30 PM ET

Best line: 3.5 at FanDuel (-122) · Confidence Score: 8.0 / 10

Utah State is 1-2 ATS under Bronco Mendenhall, who hasn't achieved much as a head coach. Despite almost beating Washington in a yawn-inspiring defensive showcase, this team got whooped by Idaho State and put up a goose egg against Utah. QB Grady Brosterhous has been careless with the ball, with 393 yards, 2 TDs, and 5 INTs, 3 of which were against Washington.

Troy, under Gerad Parker (the worst way to spell "Jared"), is also 1-2 ATS but has stayed nearly on the number (-0.9 average), and its defense ranks 87th to Utah State's 92nd. QB Will Crowder (aka "Goose" for the ball knowers) has been much better than Brosterhous (596 yards, 6 TDs, 1 INT).

Thesis. A pick-em-caliber game got priced with Utah State favored by 3, and both projections see it closer to even. Last week, Troy put up a good game in their loss to Missouri (ranked 19th), getting ~100 more yards in the error and stopping 9/11 third down conversions. In contrast, Utah State has no momentum after their Utah beatdown.

The numbers. SP+ projects Utah State -0.7 (favoring Troy relative to market), and FPI has it at Utah State -0.4, against a market line of -3. Dual edge of 3.4, projected margin 0.6.

What breaks this. This is a close spread, and one score could blow it. Utah State did somehow keep up with Washington, so that needs to be taken into account.

3. Clemson Tigers +1.5 vs. California Golden Bears. Fri 9/25, 10:30 PM ET

Best line: 1.5 at DraftKings (-118) · Confidence Score: 7.25 / 10

Clemson has been a disaster against the spread (1-2) with a 10.3 average miss under Dabo Swinney. This is the same coach who used to give prime Alabama the business. What a fall from grace. Despite having a punchable aura, QB Tait Reynolds has been solid (415 yards, 1 TD, 1 INT) after taking over for the booboo ridden Christopher Vizzina. If you ignore the beating they took from LSU, this team isn't that bad.

Cal is also struggling ATS (1-2) with a 9.3-point average miss under first-year coach and human lego, Tosh Lupoi. Cal's Jaron-Keawe Sagapolutele has actually produced (866 yards, 7 TDs, 2 INTs), but much of that comes from last week's game against the upset-minded Wagner Seahawks. Prior to Wagner, they got destroyed by the CEO of hair gel's UCLA and squeaked past Syracuse 21-18.

Thesis. Last week, Clemson outscored, out-threw, and out-rushed Belichick's crew, tits and all. In contrast, Cal only has a tight win against Syracuse, a game in which they went 3/13 on third down conversions, and it's not even cold in New York yet. What keeps Clemson live here is the defense, ranked 27th nationally against Cal's 74th, a gap both models weight heavily. Dabo also has to lock in at some point before facing Miami, which is something every man has had to do in his liftime.

The numbers. SP+ has Clemson favored by 3.9, and FPI by 5.2, against a market line of 0. Dual edge of 3.9, projected margin 4.5.

What breaks this. Clemson looked really bad against LSU, especially considering LSU's step down in performance in following games, most notably against Ole Miss. Cal is also not a bad team, and this is a very close spread that could be decided by a late game field goal.

Bonus

Moneyline upset of the week

Colorado State Rams ML +440 vs. UTSA Roadrunners, Sat 9/26, 12:00 PM ET. Best line: +440 (best of book).

New coach Jim Mora inherits a Colorado State team that has actually covered well, 2-1 ATS over its last three with an 11.1-point average cover, behind QB Hauss Hejny (619 yards, 10 TDs, 1 INT). You read that right, 10 TDs.

UTSA, under longtime coach Jeff Traylor, is 3-0 ATS over the same stretch and has its own efficient passer in Owen McCown (559 yards, 4 TDs, 1 INT),

The model's win probability for Colorado State (32.6%) is nearly double the market's implied probability (18.5%) at this price. That is the trade here.

What breaks this. UTSA's own ATS run (3-0, plus-4.3 average) and its higher SP+ ranking (74th to Colorado State's 96th) are both reasons why this is a risky one. Also, Hauss is doubtful and it will come down to a gametime decision whether he plays.

The Do Your Job parlay

South Florida (-1000), New Mexico (-450), Houston (-1050), Fresno State (-550), Penn State (-370), Virginia Tech (-575), Oregon State (-430), Florida Atlantic (-500).

Combined odds: +284.

Every leg here comes from a game the board already flagged as a projected blowout. Model win probabilities range from 81 to 91 percent, which is exactly why these favorites were passed over as spread plays but still profile as safe moneyline holds.

The card

Bonus (moneyline)

Graded against the number posted at report time.