Week 3 college football picks, predictions: My 3 best bets for Week 3

Yet again we’ve made it to the weekend. After a chaotic stretch of line movement and injury reports, it’s time to reset, review the landscape and fire on a couple more positions to round out the Week 3 college football betting card.

Injury watchThe quarterback carousel is already spinning across the country, and the ripple effects are hitting markets quickly. Three QBs I’ve been keeping close track of:

Ole Miss QB Austin Simmons has been in a walking boot all week and hasn’t practiced. Officially “probable,” but message boards suggest it’s closer to a coin flip.

Texas A&M QB Marcel Reed should be fine. Postgame pressers and market movement confirm as much, with Aggies money trimming the spread from Notre Dame -7 to -6.5 and the total creeping up to 50.5.

Injuries help paint a picture of why the market is moving and can help grab quality positions if the news is a bit over-exaggerated or the market moves too far on a player’s impact. For my early-week bets, you can check out the Monday article. Looking to use some of the information that the market has given us throughout the week, let’s turn to new positions.

Odds courtesy of BetMGM.

Arkansas at No. 17 Ole Miss (-6.5, 63)The Simmons situation at QB looms large here. If he plays at less than 100%, Ole Miss’ tempo (the core of its offense) takes a serious hit — and will likely mean fewer snaps, more conservative pacing and fewer explosive sequences. Looking back at last year’s matchup, the total opened at 57 and closed at 54.5, despite featuring a first-round QB and two NFL-bound receivers for Ole Miss. Yes, Ole Miss also lost a ton of defensive talent, but this year’s line is nearly 10 points higher, which feels inflated given the downgrade in offensive personnel and Simmons’ uncertain health.

Bet: Arkansas-Ole Miss Under 63

No. 13 Oklahoma (-22, 51.5) at TempleJohn Mateer has surged into the Heisman conversation, and the transition with coach Ben Arbuckle running the offense looks seamless. Arbuckle’s short-term history as a heavy favorite is telling: his teams rarely let off the gas. In their two matchups as favorites of 20 or more last season, Washington State scored 70 and 49 points under the Arbuckle and Mateer campaign. Temple, meanwhile, hired offensive coordinator Tyler Walker, fresh off leading Montana State to the nation’s top scoring offense at the FCS level. That philosophy is already seeping into their approach, making them a live ‘dog to put up points.

Last year’s matchup closed at 57.5 with less offensive optimism on both sides. Dropping this year’s total by nearly a touchdown doesn’t make sense given the upgrades.

Bet: Oklahoma-Temple Over 51.5

Duke at Tulane (-1.5, 54)This matchup is worth highlighting because of how the market handled it. Tulane opened as nearly a 4-point favorite following Duke’s blowout loss to Illinois. Then came the wave: Blue Devils money poured in, flipping Duke all the way to a 1.5-point favorite before resistance showed and the line settled back with them as a slight underdog.

[Check out all of Yahoo’s sports betting content here in our new betting hub]

So why would bettors back Duke fresh off a 45-19 loss? EPA per play is the answer. Despite the scoreboard, Duke actually graded out better in estimated points per attempt, suggesting the Blue Devils were far more competitive than the final margin showed. If that game were simulated multiple times, advanced metrics suggest Duke would come out on top more often than not.

For sharp bettors, that meant the opener was inflated by the optics of a lopsided box score, giving them a chance to “buy low” on the Blue Devils before the market corrected.

Bet: Duke-Tulane Under 54

Final thoughtsWeek 3 is shaping up as the first true “injury week” of the season, and the betting market has responded accordingly. For handicappers, the edge lies in balancing those adjustments with historical comps that the public market often overlooks. Whether Simmons limps out for Ole Miss or Mateer continues his Heisman surge, the value comes from spotting where the number simply doesn’t line up with reality.