This is the fourth entry in a season-long watching brief comparing SharpModels' own ratings against the market, game by game. Week 3 went clearly to the model: it picked 11 of 16 winners to the market's 8 and was closer on average. All three games where the two sides backed different teams went the model's way (Browns over Panthers, Steelers over Bengals, Broncos over Rams), and the model was also much closer on Chicago's 27–7 win over Philadelphia, a game the last entry expected the market to win because of Caleb Williams' injury. Home teams won 11 of 16, after the market had spent the week discounting home field.
A gap between model and market is worth chasing, not noise to shrug off. SharpModels' ratings are built purely from what happens on the field, so unlike the market they carry no offseason or injury news, only real results. Model probabilities are taken from the SharpModels matchups page; market probabilities are two-sided prices normalised to sum to 100. Away team is shown first, and diff is model minus market on the away side, so a negative number means the market is higher on the away team than the model is.
Week 4 Slate
| Game | Model away | Model home | Market away | Market home | Diff (away) |
|---|---|---|---|---|---|
| Steelers@ Browns | 57% | 43% | 57% | 43% | 0 |
| Coltsvs Commanders | 50% | 50% | 64% | 36% | −14 |
| Cardinals@ Giants | 40% | 60% | 54% | 46% | −14 |
| Titans@ Ravens | 13% | 87% | 15% | 85% | −2 |
| Patriots@ Bills | 29% | 71% | 27% | 73% | +2 |
| Jets@ Bears | 19% | 81% | 38% | 62% | −19 |
| Jaguars@ Bengals | 52% | 48% | 43% | 57% | +9 |
| Cowboys@ Texans | 30% | 70% | 44% | 56% | −14 |
| Packers@ Buccaneers | 50% | 50% | 63% | 37% | −13 |
| Rams@ Eagles | 44% | 56% | 60% | 40% | −16 |
| Dolphins@ Vikings | 18% | 82% | 17% | 83% | +1 |
| Broncos@ 49ers | 43% | 57% | 43% | 57% | 0 |
| Chiefs@ Raiders | 64% | 36% | 65% | 35% | −1 |
| Chargers@ Seahawks | 17% | 83% | 26% | 74% | −9 |
| Lions@ Panthers | 65% | 35% | 64% | 36% | +1 |
| Falcons@ Saints | 46% | 54% | 42% | 58% | +4 |
Mean absolute deviation is 7.4 points and the median is 6.5, with 8 of 16 games disagreeing by more than 5. The signed mean is −5.3, and the market is again higher on the away team than the model in 9 of 16 games. This week the gaps have a clearer cause than home field. Four of the six widest disagreements involve a team without its starting quarterback: Chicago (Caleb Williams), Washington (Jayden Daniels), the Giants (Jaxson Dart) and Tampa Bay (Baker Mayfield). Elo only reads final scores, so it cannot see who is under centre. Chicago's win with Case Keenum last week is a reminder that the market's adjustment for a missing quarterback can be too big.
Match by match
Steelers @ Browns Thu, Oct 1, 2026 · 8:15 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Steelers | 57% | 57% | 0 |
| Browns | 43% | 43% | 0 |
Both teams are 2–1 after close home wins in Week 3: Pittsburgh edged Cincinnati 30–27 and Cleveland held off Carolina 21–18, each worth 12 rating points. Pittsburgh (1524, 15th) is rated well clear of Cleveland (1404, 25th), enough to overcome the Browns' home edge, and model and market land on exactly the same number: Steelers 57%.
Colts vs Commanders Sun, Oct 4, 2026 · 9:30 AM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Colts | 50% | 64% | −14 |
| Commanders | 50% | 36% | +14 |
Played at a neutral site, the Tottenham Hotspur Stadium in London. Both teams won at home in Week 3 as underdogs: Indianapolis beat Houston 19–17 as a 40% shot (up 13 to 1432, 22nd), and Washington pulled the upset of the week, beating Seattle 33–31 as a 23% underdog (up 17 to 1434, 21st). Two ratings two points apart at a neutral venue make an exact coin flip on the model. The market makes the Colts a 64% favourite, and the likely reason is under centre: Jayden Daniels dislocated his elbow in the Week 2 loss at Dallas, and while he has travelled to London and is practising, Marcus Mariota is still expected to start. The rating has no way of knowing that.
Cardinals @ Giants Sun, Oct 4, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Cardinals | 40% | 54% | −14 |
| Giants | 60% | 46% | +14 |
Arizona gave San Francisco a real game in Week 3, losing 30–36 on the road as a 19% underdog (down just 7 to 1374, 30th). New York beat Tennessee 12–7 at home (up 10 to 1399, 27th). The Giants are rated slightly higher and are at home, so the model makes them a 60% favourite. The market flips it and makes Arizona a 54% road favourite. The Giants' quarterback situation explains it: Jaxson Dart sprained his MCL in the Week 2 loss to the Rams and could miss much longer, leaving Jameis Winston to start, and pass rusher Brian Burns has reportedly suffered what looked like a torn ACL. Elo reads the Giants as the team that has gone 2–1; the market is pricing the team that takes the field on Sunday.
Titans @ Ravens Sun, Oct 4, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Titans | 13% | 15% | −2 |
| Ravens | 87% | 85% | +2 |
Tennessee is 0–3 after a 7–12 loss at the Giants and remains the league's lowest-rated team (1246, 32nd). Baltimore beat Dallas 34–31 in Rio (up 10 to 1578, 11th). Close to exact agreement: Ravens around 85–87% either way.
Patriots @ Bills Sun, Oct 4, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Patriots | 29% | 27% | +2 |
| Bills | 71% | 73% | −2 |
New England was routed 6–35 at Jacksonville (down 33 to 1578, 12th), and Buffalo moved to 3–0 by beating the Chargers 24–16 at home (up 9 to 1696, 2nd). Model and market agree almost exactly: Bills by roughly 71–73%.
Jets @ Bears Sun, Oct 4, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Jets | 19% | 38% | −19 |
| Bears | 81% | 62% | +19 |
The widest gap of the week, and a rematch of last week's argument. Before Chicago played Philadelphia, this watching brief suggested Caleb Williams' hamstring injury was probably most of the market's 14-point edge over the model. The Bears then started third-stringer Case Keenum and won 27–7, a result worth 33 points that lifted them to 1559 (13th). The model had Philadelphia as a narrow favourite too, but it was much closer than the market. Williams has since been diagnosed with a Grade 2 hamstring strain, reported at three to four weeks, so Chicago will start Keenum or, if he clears the concussion protocol, Tyson Bagent. The Jets lost 24–31 at Detroit and sit 31st at 1324. The model makes Chicago an 81% favourite; the market has the Bears at just 62%. The market is still pricing the missing quarterback, and the model now has one piece of evidence that the Bears can win without him.
Jaguars @ Bengals Sun, Oct 4, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Jaguars | 52% | 43% | +9 |
| Bengals | 48% | 57% | −9 |
Jacksonville produced the most one-sided result of Week 3, beating New England 35–6 at home to jump 33 points to 1596 (7th). Cincinnati lost 27–30 at Pittsburgh (down 12 to 1520, 16th). The Jaguars' rating lead is big enough to outweigh Cincinnati's home edge, so the model leans Jacksonville at 52%. The market goes the other way, Bengals 57%, one of only a few games this week where the model is higher on the road team than the market is.
Cowboys @ Texans Sun, Oct 4, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Cowboys | 30% | 44% | −14 |
| Texans | 70% | 56% | +14 |
Houston is 0–3 after a 17–19 loss at Indianapolis as a 60% favourite, and has fallen 70 points in three weeks, the second-biggest drop in the league. Because the Texans started from a high base (1625), they are still rated 14th at 1554. Dallas lost 31–34 to Baltimore in Rio (down 10 to 1441, 19th). The model still trusts Houston's underlying rating and makes the Texans a 70% home favourite; the market has them at only 56%. Part of that is personnel: Houston's offense has struggled in the absence of top receiver Nico Collins (hamstring), though he is reported to have a decent chance of returning this week.
Packers @ Buccaneers Sun, Oct 4, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Packers | 50% | 63% | −13 |
| Buccaneers | 50% | 37% | +13 |
Green Bay suffered the worst result of Week 3, beaten 14–35 at home by Atlanta as a 73% favourite, a 45-point fall to 1492 (17th). Tampa Bay lost 16–23 at home to Minnesota and is 0–3 (down 16 to 1431, 23rd). Green Bay's 61-point rating lead is almost exactly cancelled out by home field, so the model has this at 50–50. The market makes Green Bay a 63% favourite because Baker Mayfield dislocated his right thumb in that Vikings game and is out for at least three weeks, leaving rookie Jalon Daniels to start. A starting quarterback swap is exactly the kind of news an Elo model cannot see.
Rams @ Eagles Sun, Oct 4, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Rams | 44% | 60% | −16 |
| Eagles | 56% | 40% | +16 |
Philadelphia's 7–27 loss at Chicago, to a third-string quarterback, cost the Eagles 33 points and dropped them to 1588 (8th). Jalen Hurts also took a big hit in that game, though he returned. Los Angeles lost 26–30 at Denver (down 14 to 1602, 6th). The Rams are rated 14 points higher, less than the home edge, so the model makes Philadelphia a 56% home favourite. The market flips it, Rams 60%, despite the Rams reportedly being without cornerback Jaylen Watson and tight end Terrance Ferguson. This one looks more like the market marking the Eagles down hard for Monday night than any single piece of injury news.
Dolphins @ Vikings Sun, Oct 4, 2026 · 4:05 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Dolphins | 18% | 17% | +1 |
| Vikings | 82% | 83% | −1 |
Miami is 0–3 after a 10–24 home loss to Kansas City (down 20 to 1380, 29th). Minnesota is 3–0 after a 23–16 win at Tampa Bay (up 16 to 1626, 4th). Model and market agree almost exactly: Vikings around 82–83%.
Broncos @ 49ers Sun, Oct 4, 2026 · 4:25 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Broncos | 43% | 43% | 0 |
| 49ers | 57% | 57% | 0 |
Two top-five teams. Denver beat the Rams 30–26 on Sunday night (up 14 to 1629, 3rd), and San Francisco is 3–0 after holding off Arizona 36–30 (up 7 to 1625, 5th). Four rating points separate them, so the home edge decides it on the model, and model and market land on exactly the same number: 49ers 57%.
Chiefs @ Raiders Sun, Oct 4, 2026 · 4:25 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Chiefs | 64% | 65% | −1 |
| Raiders | 36% | 35% | +1 |
Last week's widest gap is this week's agreement. The model made Kansas City only a 62% favourite at Miami while the market had the Chiefs at 85%; Kansas City won 24–10, and the rating rose 20 to 1587 (9th), up 62 over three weeks. Las Vegas is also 3–0 after winning 35–27 at New Orleans as a 35% underdog. The Raiders have gained 99 points in three weeks, the most in the league, but started so low that they are still only 26th at 1400. Model and market now agree: Chiefs 64–65% on the road.
Chargers @ Seahawks Sun, Oct 4, 2026 · 4:25 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Chargers | 17% | 26% | −9 |
| Seahawks | 83% | 74% | +9 |
Los Angeles is 0–3 after a 16–24 loss at Buffalo and has lost 86 rating points in three weeks, the most in the league (now 1447, 18th). Seattle was upset 31–33 at Washington as a 77% favourite (down 17 to 1713) but is still the top-rated team in football. The model makes Seattle an 83% home favourite; the market is a little more generous to the Chargers at 26%.
Lions @ Panthers Sun, Oct 4, 2026 · 8:20 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Lions | 65% | 64% | +1 |
| Panthers | 35% | 36% | −1 |
Detroit beat the Jets 31–24 at home (up 8 to 1585, 10th), and Carolina lost 18–21 at Cleveland (down 12 to 1393, 28th). Model and market agree to within a point: Lions around 65% on the road.
Falcons @ Saints Mon, Oct 5, 2026 · 8:15 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Falcons | 46% | 42% | +4 |
| Saints | 54% | 58% | −4 |
Atlanta produced the biggest gain of Week 3, winning 35–14 at Green Bay as a 27% underdog. That was worth 45 points, which almost exactly cancelled out the 46 it lost to Carolina in Week 2, and the Falcons are back to 1435 (20th). New Orleans lost 27–35 at home to Las Vegas (down 28 to 1406, 24th). Atlanta is now rated higher, but not by enough to overcome the Saints' home edge, so the model makes New Orleans a 54% favourite. The market is a little firmer on the Saints at 58%.