AI-powered insights from our model. Choose a matchup to view analysis.
For Carolina–Atlanta, Cooper Rush has some notable modeled price gaps, but the supplied prop rows don’t label the stat categories, so I wouldn’t assume what those lines measure. His listed under 0.5 at +185 has a 50.0% model probability versus 35.1% from the market, a positive 1,491-basis-point edge across six books that favors the under. His under 29.5 at -108 shows a 63.6% model probability versus 51.9% from the market, a positive 1,171-basis-point edge across four books. Across the labeled markets, receptions lean toward overs by angle count, 27–17, while rushing yards lean toward unders, 17–9; those counts aren’t confidence ratings. These are experimental estimates, not guarantees, and I’d want the missing stat labels confirmed before treating either Rush entry as an actionable pick.
For Saints–Ravens, Alvin Kamara over 2.5 at +150 is one angle worth a look: the model estimates 54.8% versus the market’s 40.0%, with a reported positive edge of 1,475 basis points across four books. Derrick Henry over 1.5 at +182 has a similar disagreement—50.0% from the model versus 35.5% from the market, a positive 1,453-basis-point edge across five books. One important catch: the supplied data doesn’t identify the prop category for either line, so you’d need to confirm that before treating either as an actionable pick. Across the listed markets, rushing yards lean toward overs, while passing attempts and completions lean toward unders; passing yards and receiving yards are evenly split. These are experimental estimates, so the gaps are interesting signals—not established accuracy or guarantees.
For Vikings–Bears, Kalif Raymond over 2.5 at +133 has the largest supplied edge: the model gives it a 57.9% chance versus the market’s 42.9%, a reported +1,497 basis points across five books. Carson Wentz over 1.5 at +148 is another model-backed over, with a 53.3% estimate versus 40.3% from the market and a +1,301-basis-point edge across six books. One important catch: the supplied player rows don’t identify the stat attached to either line, so you’ll need to confirm that before acting on them. These are sizable model–market disagreements, but the estimates are experimental—not proven win rates or guarantees.
For Bengals–Texans, Tee Higgins under 3.5 at +140 is worth a look: the model gives that under a 54.1% chance versus the market’s 41.7%, with a reported +1,242-basis-point edge across four books. Samaje Perine over 1.5 at +135 is another supported angle, with a 54.8% model probability versus 42.6% from the market and a +1,220-basis-point edge across five books. Those positive edges favor the specific bets listed—Higgins’ under and Perine’s over—not overs in general. These are experimental estimates rather than guarantees, so I’d treat the gaps as reasons to consider those prices, not proof that either bet will win.
Deshaun Watson under 0.5 at +158 is worth a look: the model gives that under a 50.0% chance versus the market’s 38.8%, a positive edge of 1,124 basis points—about 11.2 percentage points—across six books. Bucky Irving over 2.5 at +123 is another supported angle, with a modeled 54.8% chance versus 44.8% from the market, a 991-basis-point edge—about 9.9 percentage points—in favor of the over across five books. More broadly, the supplied directions lean toward passing-touchdown and rushing-attempt unders, while receptions lean toward overs. These are experimental estimates, not guarantees: the appeal is the disagreement with the prices, not certainty that either prop will hit.
For Packers–Jets, the strongest supplied angle is Geno Smith under 30.5 at -105. The model estimates a 63.6% chance for that under versus the market’s 51.2%, a positive edge of 1,241 basis points across three books. Tucker Kraft under 3.5 at +115 also has support: the model gives the under 57.9% versus the market’s 46.5%, with a supplied edge of 1,138 basis points across five books. More broadly, the modeled passing markets lean toward unders, while receptions are more mixed—those positive edges favor the named under bets, not overs. These are experimental estimates rather than proven hit rates or guarantees, so I’d describe both as promising model disagreements, not high-confidence outcomes.
Aaron Rodgers under 0.5 at +185 has the largest supplied edge: the model gives that under a 50.0% chance versus the market’s 35.1%, a positive 1,491-basis-point edge across five books. Roman Wilson under 2.5 at +116 is another angle, with a 54.8% model probability versus 46.3% from the market—a positive 846-basis-point edge across five books, again favoring the under. One important catch: the supplied player rows don’t identify the stat attached to either line, so you’ll need to confirm the prop labels before acting on them. These are experimental estimates, not guarantees, and the probability gaps suggest potential value rather than proven predictive accuracy.
Saquon Barkley under 1.5 at +140 has the largest supplied edge: the model estimates a 54.1% chance versus the market’s 41.7%, a reported +1,242 basis points across four books. Jalen Hurts over 1.5 at +140 is close behind, with a 53.3% model probability versus 41.7% from the market and a +1,166-basis-point edge across six books. Those positive edges mean the model favors the named sides relative to their prices—not that both picks are overs. One important catch: the prop categories for these lines aren’t identified in the supplied evidence, so confirm what each line measures before treating either as a betting recommendation. These are experimental estimates, not proven win rates or guarantees, even with evidence spanning multiple books.
For Jaguars–Broncos, Evan Engram over 3.5 at +136 is worth a look: the model gives it a 54.8% chance versus the market’s 42.4%, a reported +1,238-basis-point edge across 3 books. Adam Trautman over 1.5 at +135 has a similar case, with a 54.1% model probability against 42.6% from the market and a +1,153-basis-point edge across 6 books. The broader market directions lean under on passing attempts and completions but over on passing touchdowns, while passing yards are evenly split. Those are meaningful model-versus-market disagreements, but the estimates are experimental—not proven advantages or guarantees—and neither model probability makes the over close to a sure thing.
Tre Harris under 2.5 at +133 is the clearest supplied angle: the model gives it a 54.1% chance versus the market’s 42.9%, a positive edge of 1,116 basis points across four books. Kirk Cousins under 0.5 at +190 also shows a positive edge on the under, with a 44.2% model probability versus 34.5% from the market—a 972-basis-point gap across three books. The distinction is that Harris’s under is modeled as slightly more likely than not, while Cousins’s under still falls below 50% despite the favorable price comparison. More broadly, the supplied directions lean toward unders in receptions, rushing attempts, and rushing yards, while passing and receiving yards are evenly split. These are experimental estimates, not guarantees, so I’d view the probability gaps as reasons to consider those unders rather than proof they’ll hit.
For Seattle–Arizona, the two angles worth a closer look are Jacoby Brissett over 1.5 at +200 and Drew Lock under 29.5 at -105. On Brissett’s over, the model estimates a 48.2% chance versus the market’s 33.3%, with a reported positive edge of 1,488 basis points across six books—even though the model still puts it below a coin flip. On Lock’s under, it estimates 63.6% versus 51.2%, a positive edge of 1,241 basis points across four books; that edge favors the under, not the over. The broader modeled directions lean toward passing-touchdown overs and passing-attempt unders, while passing yards are evenly split. These are experimental estimates, so I’d treat them as interesting model–market disagreements rather than reliable predictions or guarantees.
For Washington–Dallas, the two strongest supplied angles are unders. Dak Prescott under 1.5 at +159 has a 50.0% model probability versus the market’s 38.6%, a reported +1,138-basis-point edge across six books. Javonte Williams under 17.5 at -101 has a 59.5% model probability versus 50.2% from the market, a reported +921-basis-point edge across four books. Those positive edges favor the named under bets relative to their prices; they don’t signal an over or make either outcome a sure thing. These are experimental estimates, so treat them as potential pricing disagreements rather than proven advantages or guarantees.
The two biggest supplied price disagreements are Malik Willis under 0.5 at +130 and De’Von Achane under 3.5 at +127, although the data doesn’t identify the stat behind either line, so you’d want to confirm that first. For Willis, the model gives the under a 55.8% chance versus the market’s 43.5%, with a reported edge of 1,231 basis points—about 12.3 percentage points—across seven books. For Achane, it’s 54.8% versus 44.1%, with a reported 1,070-basis-point edge, or 10.7 percentage points, across three books. Those positive edges favor the named unders; they don’t signal an over, and no specific best-priced book is supplied. These are experimental estimates rather than guarantees, so the disagreements are worth a closer look, not a reason to call either bet a sure thing.
Noah Gray over 1.5 at +148 is worth a look: the model gives it a 54.8% chance versus the market’s 40.3%, with a reported 14.43-percentage-point edge across five books. Daniel Jones under 31.5 at -110 has a higher modeled probability at 63.6% versus 52.4% from the market, with an 11.25-point edge across three books. Jones over 1.5 at +180 also has support—46.7% from the model versus 35.7% from the market, a 10.95-point edge across six books—but the model still puts that outcome below a coin flip. Those positive edges describe disagreement with the market price, not an “over” direction, which is why Jones’s under can have a positive edge too. These are experimental estimates, not guarantees, and the book counts alone don’t establish how reliable the model is.
Jaxson Dart over 1.5 at +155 has the largest supplied disagreement: the model gives that over a 53.3% chance versus the market’s 39.2%, a reported +1,411-basis-point edge across seven books. Cam Skattebo under 1.5 at +118 is another angle, with a 54.2% model probability versus 45.9% from the market and a reported +833-basis-point edge across six books. That positive edge supports Skattebo’s **under**, not an over—the sign describes the modeled advantage, not the bet direction. The supplied player lines don’t identify their stat categories, so you’d need to confirm those markets before acting on either angle. These are experimental estimates worth investigating, not established probabilities or guarantees.