NFL Predictions Week 3: The Wrong Team Is Favored In Denver
By Jake Hari
August 12, 2026 · Updated September 23, 2026
Updated September 23, 2026 · 11 min read · by Jake Hari
The Rams are the road favorite at Denver on Sunday night, and six AI models that priced every Week 3 game without seeing a price all say the wrong team is favored. Denver sits at 44.5 cents on Kalshi and the six models average 62%, with every one of them between 60 and 63. It is the widest gap on the board, and the second widest runs the other way: New Orleans is 62.5 cents at home against the Raiders and the panel calls that game a coin flip. Through two weeks the market has been closer to the result on 22 of 32 games and the panel on 10, so this is not a claim that the models are smarter than the money. It is a map of the four games where the two disagree enough to be worth your Sunday.
New to event markets? Prediction Markets For DFS Players covers the mechanics in plain language; this page is the weekly game-by-game layer on top of it. The full OddsShopper edition, with the exchange comparison and the dated injury record, is AI NFL Predictions Week 3; this page is the Stokastic.com side of that pair.
The Board: All 16 Games
Sixteen games, home team in bold, ordered by the size of the gap so the disagreements come first. "Kalshi: home" is the Tuesday price of the home team's winner contract in cents, which reads as a percentage. "Panel: home" is the mean of six estimates, one per model, each made from the same fetched card without seeing a price. A positive gap means the models like the home team more than the market does. Gemini did not return a board this week.
| Game (Away @ Home) | Kickoff (ET) | Kalshi: home (Tue) | Panel: home | Gap |
|---|---|---|---|---|
| Rams @ Broncos | Sun 8:20 p.m. | 44.5¢ | 62.0% | +17.5 |
| Raiders @ Saints | Sun 4:25 p.m. | 62.5¢ | 47.7% | -14.8 |
| Panthers @ Browns | Sun 1:00 p.m. | 40.5¢ | 49.2% | +8.7 |
| Vikings @ Buccaneers | Sun 4:05 p.m. | 46.5¢ | 38.5% | -8.0 |
| Chiefs @ Dolphins | Sun 1:00 p.m. | 13.5¢ | 20.2% | +6.7 |
| Patriots @ Jaguars | Sun 1:00 p.m. | 58.5¢ | 52.3% | -6.2 |
| Titans @ Giants | Sun 1:00 p.m. | 62.5¢ | 57.3% | -5.2 |
| Eagles @ Bears | Mon 8:15 p.m. | 36.5¢ | 31.7% | -4.8 |
| Bengals @ Steelers | Sun 1:00 p.m. | 37.5¢ | 42.0% | +4.5 |
| Chargers @ Bills | Sun 1:00 p.m. | 74.5¢ | 77.5% | +3.0 |
| Jets @ Lions | Sun 1:00 p.m. | 73.5¢ | 70.7% | -2.8 |
| Ravens @ Cowboys | Sun 4:25 p.m. | 41.5¢ | 43.8% | +2.3 |
| Falcons @ Packers | Thu 8:15 p.m. | 71.5¢ | 70.0% | -1.5 |
| Texans @ Colts | Sun 1:00 p.m. | 42.5¢ | 43.8% | +1.3 |
| Cardinals @ 49ers | Sun 4:05 p.m. | 78.5¢ | 77.3% | -1.2 |
| Seahawks @ Commanders | Sun 1:00 p.m. | 27.5¢ | 26.5% | -1.0 |
Prices as of Tuesday, September 22, 2026, 7:15 a.m. ET, Kalshi asks before fees. The panel's numbers were generated the same morning, price-blind, and do not move; the market's do, and every row grades once the game settles. These are estimates, not predictions of fact and not financial advice.
Every Model's Number
The blend hides the argument, so here is each model's probability that the home team wins, with the Tuesday price beside it.
| Game | Claude Fable | Claude Opus | Claude Sonnet | ChatGPT | Grok | Kimi | Blend | Kalshi |
|---|---|---|---|---|---|---|---|---|
| ATL @ GB | 71% | 74% | 72% | 68% | 67% | 68% | 70.0% | 71.5¢ |
| CAR @ CLE | 50% | 51% | 52% | 47% | 45% | 50% | 49.2% | 40.5¢ |
| CIN @ PIT | 43% | 42% | 42% | 42% | 41% | 42% | 42.0% | 37.5¢ |
| HOU @ IND | 48% | 47% | 43% | 43% | 40% | 42% | 43.8% | 42.5¢ |
| KC @ MIA | 19% | 20% | 22% | 22% | 16% | 22% | 20.2% | 13.5¢ |
| LAC @ BUF | 79% | 82% | 78% | 75% | 73% | 78% | 77.5% | 74.5¢ |
| NE @ JAX | 54% | 52% | 48% | 52% | 53% | 55% | 52.3% | 58.5¢ |
| NYJ @ DET | 70% | 73% | 69% | 68% | 72% | 72% | 70.7% | 73.5¢ |
| TEN @ NYG | 58% | 58% | 57% | 57% | 58% | 56% | 57.3% | 62.5¢ |
| ARI @ SF | 76% | 79% | 79% | 78% | 76% | 76% | 77.3% | 78.5¢ |
| MIN @ TB | 38% | 40% | 34% | 41% | 40% | 38% | 38.5% | 46.5¢ |
| BAL @ DAL | 45% | 43% | 40% | 46% | 43% | 46% | 43.8% | 41.5¢ |
| LV @ NO | 50% | 52% | 44% | 47% | 46% | 47% | 47.7% | 62.5¢ |
| LAR @ DEN | 63% | 62% | 63% | 60% | 61% | 63% | 62.0% | 44.5¢ |
| PHI @ CHI | 30% | 33% | 32% | 34% | 28% | 33% | 31.7% | 36.5¢ |
| SEA @ WSH | 28% | 23% | 28% | 28% | 25% | 27% | 26.5% | 27.5¢ |
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $178M traded
- 2027 Pro Football Champion · $60M traded
- 2028 U.S. Presidential Election winner? · $59M traded
- 2028 Republican presidential nominee · $56M traded
- Pro Baseball Champion · $52M traded
- More tech layoffs in 2026 than in 2025? · $31M traded
Every market above links to our full AI model verdict; browse them all on the Stokastic prediction markets hub, and see how every settled call actually scored on the full graded scoreboard.
Week 2, Graded
The market won the week 10-6, and the shape of the split is the useful part. Five favorites lost, and on four of the five the panel was the side nearer the result: the Buccaneers at 78.5 cents against a 70.0% blend, the Bears at 65.5 against 59.4%, the Ravens at 78.5 against 75.9%, the Chargers at 72.5 against 66.4%. The exception was Houston, where the panel had the Texans at 61.6% against a 58.5-cent price and Cincinnati won 20-6. The market earned its ten on the favorites that won big: the 49ers by 22, the Cowboys by 17, the Seahawks by 24. Kalshi is the price on the OddsShopper board before kickoff (Tuesday's for Thursday night, Sunday 7 a.m. ET for the rest); closer means nearer to the $1 or $0 the contract paid.
| Game (Away @ Home) | Final | Kalshi: home | Panel: home | Closer |
|---|---|---|---|---|
| Lions @ Bills | BUF 41-31 | 67¢ → $1 | 66.3% | market closer by 0.7 |
| Panthers @ Falcons | CAR 34-3 | 42.5¢ → $0 | 43.1% | market closer by 0.6 |
| Bengals @ Texans | CIN 20-6 | 58.5¢ → $0 | 61.6% | market closer by 3.1 |
| Browns @ Buccaneers | CLE 23-19 | 78.5¢ → $0 | 70.0% | panel closer by 8.5 |
| Packers @ Jets | GB 20-17 OT | 38.5¢ → $0 | 37.3% | panel closer by 1.2 |
| Vikings @ Bears | MIN 9-3 | 65.5¢ → $0 | 59.4% | panel closer by 6.1 |
| Saints @ Ravens | NO 24-17 | 78.5¢ → $0 | 75.9% | panel closer by 2.6 |
| Eagles @ Titans | PHI 24-20 | 24.5¢ → $0 | 23.6% | panel closer by 0.9 |
| Steelers @ Patriots | NE 20-3 | 68.5¢ → $1 | 64.0% | market closer by 4.5 |
| Jaguars @ Broncos | DEN 20-13 | 58.5¢ → $1 | 50.6% | market closer by 7.9 |
| Raiders @ Chargers | LV 26-14 | 72.5¢ → $0 | 66.4% | panel closer by 6.1 |
| Dolphins @ 49ers | SF 35-13 | 89.5¢ → $1 | 77.4% | market closer by 12.1 |
| Seahawks @ Cardinals | SEA 31-7 | 34.5¢ → $0 | 43.4% | market closer by 8.9 |
| Commanders @ Cowboys | DAL 37-20 | 65.5¢ → $1 | 53.9% | market closer by 11.6 |
| Colts @ Chiefs | KC 33-30 OT | 71.5¢ → $1 | 68.1% | market closer by 3.4 |
| Giants @ Rams | LAR 28-6 | 74¢ → $1 | 69.4% | market closer by 4.6 |
Week 1, priced by the same method, went 12-4 the market's way, with four favorites losing (the Rams at 64.5 cents, the Titans at 51.5, the Chargers at 80.5 and the Cowboys at 60.5) and the panel's widest miss at Broncos at Chiefs, where it had Denver at 54.5% and Kansas City won 31-10. Two weeks in, the pattern is the one a price-blind panel produces: closer when a favorite falls, farther when a favorite rolls, because it will not pay 89 cents for anyone.
Rams At Broncos: The Favorite Is On The Wrong Sideline
Sunday night is the game. Denver sits at 44.5 cents on Kalshi, which makes the Rams a road favorite, and the six models have the Broncos at 60, 61, 62, 63, 63 and 63. That is a three-point band around 62% on a team the market has as an underdog, and it is the widest gap on the board by a distance.
The card the models priced from explains the number. Denver went 14-3 last season as the AFC's top seed, beat Jacksonville 20-13 at home in Week 2, and Bo Nix is not on the injury report. The Rams' report carries Puka Nacua, Jordan Whittington and Kamren Kinchens. What the market is paying for is Monday night, when the Rams beat the Giants 28-6 and the panel was 4.6 under the price. The panel is paying for the altitude and the healthier roster, and it is not moved by a home win over New York. Size the Denver side with the Week 1 miss in mind, when the panel had the Broncos at 54.5% in Kansas City and they lost by 21. By Wednesday 1:40 a.m. ET the Denver ask had slipped half a cent to 44; the market has not blinked.
Raiders At Saints: New Orleans Is Priced Like A Team It Is Not
The Saints are 62.5 cents at home and the panel has them at 47.7%, with the highest of the six at 52 and the lowest at 44. Nobody on the panel makes New Orleans a favorite of any size.
Both teams come in off a road win, and that is the whole argument for the price. New Orleans won at Baltimore 24-17 against a 78.5-cent home favorite; Las Vegas won at the Chargers 26-14 against a 72.5-cent one. The panel was the closer side on both. Then it looked at the two rosters. Las Vegas has Kirk Cousins at quarterback. The Saints' quarterback room is Spencer Rattler, Tyler Shough and Zach Wilson. These teams finished 6-11 and 3-14 last season, and one road win each does not separate them by 25 cents. A coin-flip blend is not a Raiders pick; it is a statement that 62.5 cents is too much to pay for the Saints. Wednesday's ask had come in to 61 cents, the biggest move of the four story games, and still 13 points above the panel.
Panthers At Browns: A Home Underdog The Panel Has At Even
Cleveland is a 40.5-cent home underdog and the panel has the Browns at 49.2%, four models at 50 or above and Grok the low at 45.
This is less a case for Cleveland than a case against the Carolina price. The Browns went 5-12 last season and just won at Tampa Bay 23-19 against a 78.5-cent home favorite, the row where the panel beat the market by the most in Week 2, 8.5 points. Carolina beat Atlanta 34-3, which is the result that has the Panthers favored on the road, and the card has Jonathon Brooks doubtful. A team that finished 8-9 last year favored on the road off one blowout is a price the models will not match. They land at a coin flip, and the Wednesday ask had moved a cent and a half toward them, to 42.
Vikings At Buccaneers: The Gap That Runs Against The Home Team
Tampa Bay is 46.5 cents at home and the six models have the Buccaneers between 34 and 41, a 38.5% blend, eight points under the price.
Minnesota just won at Chicago 9-3 against a 65.5-cent home favorite, one of the panel's four correct upset calls in Week 2. Tampa Bay lost at home to Cleveland, a team that went 5-12 last season, as a 78.5-cent favorite. The market has taken the Buccaneers from a heavy favorite to a slight underdog in a week; the panel has taken them further. The Tampa Bay ask sat at 47 cents on Wednesday morning, unchanged in any way that matters.
What This Board Means For Your DFS Lineups
Win probability is not a projection, but the shape of a game tells you what kind of lineup it feeds.
The panel's fades are the late-throwing games. A 62-cent favorite is a team expected to be leading in the fourth quarter, which is where a favored back gets the clock-killing carries. A coin flip is a game where both offenses are still throwing late. The panel calls Raiders at Saints and Panthers at Browns coin flips, so on its read those are two games where the trailing team's passing game stays live into the fourth quarter, and whichever side the field builds around, the other side's quarterback stack is the same game at lower ownership. NFL DFS leverage is the long version, NFL DFS stacking is how the pieces go in together, and how to use NFL DFS Sims is the settings walk-through.
The heavy favorites are where the two sides agree. The 49ers at 78.5 cents against 77.3%, the Bills at 74.5 against 77.5%, the Lions at 73.5 against 70.7%, the Packers at 71.5 against 70.0%: four games, none more than three points apart, and the game-script spots where the favored offense's touches are the safest thing on the slate.
Check the field last. NFL DFS projections, ownership and stacks in the DataHub put projected ownership next to every projection in these games, and you can try the Sims free to see how often a build around one of the coin-flip games wins its contest. Week 3 is live in the NFL Sims.
The Bottom Line
The market and the models agree on twelve of the sixteen games, which is the normal outcome and a decent argument that both are working. The four they part on are Denver at 44.5 cents against a 62% blend, New Orleans at 62.5 against 47.7%, Cleveland at 40.5 against 49.2%, and Tampa Bay at 46.5 against 38.5%. Three of the four are the panel refusing to pay for a favorite off one good result, and Sunday night decides whether that refusal was right on the biggest one.
This page is rebuilt with the next slate once these games settle, with every row above graded; the season-long version of the same comparison lives on NFL win projections 2026 and the Sims vs Kalshi ledger. Anyone can publish predictions; the useful thing is a record.
Kalshi is a CFTC-regulated event-contract exchange, not a sportsbook. These are estimates, not predictions of fact and not financial advice.
Verdicts generated Tuesday, September 22, 2026, price-blind. Panel: Claude Fable, Claude Opus, Claude Sonnet, ChatGPT, Grok and Kimi; the blend is the simple mean of the six. Gemini did not return a board. Models are frequently wrong; the market price reflects real traders' money.
More on this: NFL MVP Odds 2026 · NFL Win Projections 2026: The Teams That Moved Before Week 1 · NFL Head Coach Changes: The Market On Who's Out, And What It Does To Fantasy Values · NFL Playoff Odds: Stokastic Sims Vs Kalshi Prices · NFL Trade Deadline Odds: The Live Player-Movement Board