This is the third entry in a season-long watching brief comparing SharpModels' own ratings against the market, game by game. Week 2 went to the market: it picked 11 of 16 winners to the model's 10 and was closer on average, helped most by Carolina's 34–3 win at Atlanta, the game the last entry flagged as the widest disagreement. Every team now has two results behind it, and the ratings below reflect both.
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 3 Slate
| Game | Model away | Model home | Market away | Market home | Diff (away) |
|---|---|---|---|---|---|
| Falcons@ Packers | 27% | 73% | 30% | 70% | −3 |
| Chargers@ Bills | 19% | 81% | 25% | 75% | −6 |
| Panthers@ Browns | 44% | 56% | 58% | 42% | −14 |
| Bengals@ Steelers | 45% | 55% | 63% | 37% | −18 |
| Jets@ Lions | 18% | 82% | 27% | 73% | −9 |
| Texans@ Colts | 60% | 40% | 58% | 42% | +2 |
| Patriots@ Jaguars | 48% | 52% | 41% | 59% | +7 |
| Chiefs@ Dolphins | 62% | 38% | 85% | 15% | −23 |
| Titans@ Giants | 28% | 72% | 43% | 57% | −15 |
| Seahawks@ Commanders | 77% | 23% | 74% | 26% | +3 |
| Cardinals@ 49ers | 19% | 81% | 22% | 78% | −3 |
| Vikings@ Buccaneers | 62% | 38% | 53% | 47% | +9 |
| Ravensvs Cowboys | 64% | 36% | 63% | 37% | +1 |
| Raiders@ Saints | 35% | 65% | 40% | 60% | −5 |
| Rams@ Broncos | 43% | 57% | 56% | 44% | −13 |
| Eagles@ Bears | 54% | 46% | 68% | 32% | −14 |
Mean absolute deviation is 9.1 points and the median is 8, with 10 of 16 games disagreeing by more than 5, up from 6.8 and 7 games last week. Unlike Week 2 there is a clear direction to it: the signed mean is −6.3, and the market is higher on the away team than the model in 11 of 16 games. The model applies the same home-field edge to every game; the market is giving home teams noticeably less than that this week, perhaps because away teams won 9 of 16 games in Week 2. Three games (Panthers @ Browns, Bengals @ Steelers and Rams @ Broncos) are outright flips, with the model backing the home side and the market backing the road team.
Match by match
Falcons @ Packers Thu, Sep 24, 2026 · 8:15 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Falcons | 27% | 30% | −3 |
| Packers | 73% | 70% | +3 |
Last week's widest gap has been settled emphatically in the market's favour. The model made Atlanta a 66% home favourite over Carolina in Week 2; the market had the Panthers at 57%, and Carolina won 34–3 in Atlanta. That cost the Falcons 46 rating points, the largest fall of the week, and at 0–2 they now sit 27th at 1390. Green Bay scraped past the Jets on the road (20–17, up 10 to 1537, 14th). With the Falcons' rating finally catching up, model and market now tell the same story: Packers by roughly seven-in-ten at home.
Chargers @ Bills Sun, Sep 27, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Chargers | 19% | 25% | −6 |
| Bills | 81% | 75% | +6 |
Los Angeles has lost both of its home games by the identical score, 14–26, first to Arizona and then to Las Vegas, both times as roughly a three-to-one favourite. Those two results have taken 77 points off the rating (1533 to 1456, 18th), the biggest two-week fall in the league. Buffalo, meanwhile, beat Detroit 41–31 at home and is up 16 to 1688, second only to Seattle. The model has fully priced in the Chargers' start and makes Buffalo an 81% favourite; the market still gives Los Angeles a quarter of the win probability, six points more than the model does.
Panthers @ Browns Sun, Sep 27, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Panthers | 44% | 58% | −14 |
| Browns | 56% | 42% | +14 |
The team of the week on the ratings. Carolina's 34–3 win at Atlanta was worth 46 points, the largest gain in the league, lifting the Panthers to 1406 (24th). Cleveland also won on the road, 23–19 at Tampa Bay as a 31% underdog, and climbed 22 to 1391 (26th). The two ratings are only 15 points apart, in Carolina's favour, so the model's lean to Cleveland is almost entirely home field. The market goes the other way and makes the Panthers a 58% road favourite, one of several games this week where it gives the home side far less than the model's standard home edge.
Bengals @ Steelers Sun, Sep 27, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Bengals | 45% | 63% | −18 |
| Steelers | 55% | 37% | +18 |
Cincinnati is 2–0 after a 20–6 win at Houston as a 30% underdog that was worth 38 points and took the Bengals to 1533 (15th). Pittsburgh was held to a field goal at New England (3–20) and dropped 21 to 1511 (17th). Cincinnati is rated higher, but only by 22 points, which is less than the home-field edge the model applies, so Pittsburgh comes out a narrow 55% favourite. The market has no such hesitation: Bengals 63%, the second-widest gap on the board.
Jets @ Lions Sun, Sep 27, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Jets | 18% | 27% | −9 |
| Lions | 82% | 73% | +9 |
Both teams lost in Week 2, New York narrowly at home to Green Bay (17–20, down 10 to 1332, 31st) and Detroit in a shootout at Buffalo (31–41, down 16 to 1577, 9th). Losing at the league's second-rated team does little damage to a rating, and Detroit remains well clear of the Jets on the model's numbers. The model makes the Lions an 82% home favourite; the market is a little less sure at 73%.
Texans @ Colts Sun, Sep 27, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Texans | 60% | 58% | +2 |
| Colts | 40% | 42% | −2 |
Two 0–2 teams, but not two equally bad ones on the numbers. Houston has lost at home to Buffalo and Cincinnati, the second of those 6–20 as a 70% favourite, and has shed 58 points in two weeks. Even so, the Texans' 1567 is still 11th in the league, because they started from a high preseason base. Indianapolis lost 30–33 at Kansas City and sits 22nd at 1418. Model and market land in almost exactly the same place: Houston around 60% on the road.
Patriots @ Jaguars Sun, Sep 27, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Patriots | 48% | 41% | +7 |
| Jaguars | 52% | 59% | −7 |
New England bounced back from its opening loss at Seattle by holding Pittsburgh to a single field goal, winning 20–3 at home to climb 21 points to 1610 (7th). Jacksonville lost a close one at Denver (13–20, down 17 to 1563, 13th). The Patriots are rated 47 points higher, which the model's home-field adjustment reduces to a near coin flip (48–52). This is one of only a handful of games where the model is higher on the away team than the market: the market makes Jacksonville a 59% favourite.
Chiefs @ Dolphins Sun, Sep 27, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Chiefs | 62% | 85% | −23 |
| Dolphins | 38% | 15% | +23 |
The widest gap of the week, and the same argument as last week in a different place. Kansas City is 2–0 after edging Indianapolis 33–30 at home, but that close finish added only 8 points and the Chiefs sit 12th at 1567, still carrying much of the weight of their 2025 season. Miami is 0–2 after a 13–35 loss at San Francisco (down 14 to 1400, 25th). The model sees a 167-point gap, trimmed by home field, and makes Kansas City a 62% favourite. The market has the Chiefs at 85%, pricing them as a team with Patrick Mahomes back from last season's ACL injury rather than the team the 2025 results describe. Elo will only catch up with that view one result at a time.
Titans @ Giants Sun, Sep 27, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Titans | 28% | 43% | −15 |
| Giants | 72% | 57% | +15 |
Tennessee is 0–2 and still the league's lowest-rated team (1256, 32nd), but last week was far from a rout: the Titans lost 20–24 at home to Philadelphia as a 20% underdog, costing them just 6 points. New York was well beaten at the Rams (6–28, down 14 to 1389, 28th). The model still sees a big gap between the two and makes the Giants a 72% home favourite; the market reads this as a much closer game at 57–43, a 15-point disagreement.
Seahawks @ Commanders Sun, Sep 27, 2026 · 1:00 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Seahawks | 77% | 74% | +3 |
| Commanders | 23% | 26% | −3 |
Seattle remains the best-rated team in football after winning 31–7 at Arizona, up 14 to 1730. Washington is 0–2 after losing 20–37 at Dallas (down 26 to 1417, 23rd). Model and market agree almost exactly: Seattle around three-to-one on the road.
Cardinals @ 49ers Sun, Sep 27, 2026 · 4:05 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Cardinals | 19% | 22% | −3 |
| 49ers | 81% | 78% | +3 |
Arizona's Week 1 upset at the Chargers looked less like a turning point after a 7–31 home loss to Seattle (down 14 to 1381, 29th). San Francisco is 2–0 after beating Miami 35–13 at home and sits 4th at 1618. Another game with close agreement: the 49ers around four-to-one either way.
Vikings @ Buccaneers Sun, Sep 27, 2026 · 4:05 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Vikings | 62% | 53% | +9 |
| Buccaneers | 38% | 47% | −9 |
Minnesota won the lowest-scoring game of Week 2, 9–3 at Chicago, and is 2–0 and up 20 to 1610 (8th). Tampa Bay is 0–2 after a 19–23 home loss to Cleveland as a 69% favourite, down 22 to 1447 (20th). The ratings gap is big enough to survive the home-field adjustment, and the model makes Minnesota a 62% road favourite. The market only has the Vikings at 53%, perhaps reading a 9–3 win as less convincing than the rating does.
Ravens vs Cowboys Sun, Sep 27, 2026 · 4:25 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Ravens | 64% | 63% | +1 |
| Cowboys | 36% | 37% | −1 |
Played at a neutral site, the Maracanã in Rio de Janeiro, so neither side gets the model's home-field edge. Baltimore was the biggest upset loser of Week 2, beaten 17–24 at home by New Orleans as a 78% favourite, which cost 32 points (down to 1568, 10th). Dallas beat Washington 37–20 at home and rose 26 to 1451 (19th). Even after the Ravens' fall, the gap between the two ratings is wide, and this is the closest agreement on the board: Baltimore 64% on the model, 63% on the market.
Raiders @ Saints Sun, Sep 27, 2026 · 4:25 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Raiders | 35% | 40% | −5 |
| Saints | 65% | 60% | +5 |
Las Vegas is 2–0 and has risen 71 points in two weeks, the biggest climb in the league, after following its Week 1 win over Miami with a 26–14 win at the Chargers as a 25% underdog. Because it started from so far down, that still only gets the Raiders to 1372 (30th). New Orleans pulled Week 2's other big road upset, 24–17 at Baltimore as a 22% underdog, and jumped 32 to 1434 (21st). The model makes the Saints a 65% home favourite; the market is slightly more willing to back the Raiders' start, at 40%.
Rams @ Broncos Sun, Sep 27, 2026 · 8:20 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Rams | 43% | 56% | −13 |
| Broncos | 57% | 44% | +13 |
The purest home-field test of the week. Both teams won at home in Week 2, Los Angeles 28–6 over the Giants and Denver 20–13 over Jacksonville, and they come in with identical ratings: 1616 each, 5th and 6th in the league. With nothing to separate them, the model's 57–43 lean to Denver is all home field. The market flips it and makes the Rams a 56% road favourite, a 13-point gap on the game where the model's view is simplest to read.
Eagles @ Bears Mon, Sep 28, 2026 · 8:15 PM ET
| Team | Model | Market | Diff |
|---|---|---|---|
| Eagles | 54% | 68% | −14 |
| Bears | 46% | 32% | +14 |
Philadelphia is 2–0 but has won both games by a combined six points, 24–22 at home to Washington and 24–20 at Tennessee, so its rating has moved just 6 points each week and sits at 1620 (3rd). Chicago lost 3–9 at home to Minnesota and dropped 20 to 1526 (16th). The Eagles are rated 94 points higher, but after home field the model only makes them a 54% favourite. The market has Philadelphia at 68%, and here there is a concrete reason the model can't see. Caleb Williams was carted off with a right hamstring injury midway through the fourth quarter of that Vikings loss, and Ben Johnson has called him week to week, likely to miss time, though not ruled out for Monday. His backup, Tyson Bagent, has reportedly been placed in the concussion protocol, which could leave Chicago starting third-stringer Case Keenum. Chicago's 1526 rating was built on games Williams started; an Elo model only reads final scores, so it has no way to mark the Bears down for who is taking the snaps. That is probably most of this 14-point gap.