Week 6 previews

Thu Oct 15, 7:15 PM CT

Seattle Seahawks at Denver Broncos

Over/under voided Oct 15

Too close to call

Vegas says Seattle by 2.5, 40.5 total points. Our model says Seattle by 3.3, 43.8 total points.

Seattle should win a low-scoring road game. Both teams are 3-2, but Seattle's defense has been one of the league's best (2nd of 32 in EPA allowed per play), and our model gives the Seahawks about a 60% chance. The model and Vegas agree on the winner and nearly on the margin. Where they split is the score: the model expects about 3 more points.

The schedule still lists Drew Lock. ESPN's depth chart has Sam Darnold at QB1 and he is not on the injury report, so we use Darnold. Bo Nix starts for Denver.

Newcomer tip: success rate allowed A play is a success when it gains enough yards to stay on schedule for that down. Seattle's defense lets opponents succeed on about 38.5% of plays, 2nd lowest in the league. For a defense, lower is better.

The locked over/under (over 40.5) was voided Oct 15 and is not a pick. On the margin the gap is 0.8 points, short of the 2 needed.

Official pick:

Sun Oct 18, 8:30 AM CT

Houston Texans at Jacksonville Jaguars

Too close to call

Vegas says Jacksonville by 2.5, 42.5 total points. Our model says Jacksonville by 3.5, 42.7 total points.

Jacksonville should win. The Jaguars are 4-1 and have allowed 70 points in five games, the second fewest in the league, while Houston is 1-4. Our model gives Jacksonville about a 60% chance with Trevor Lawrence against C.J. Stroud, and Vegas sees it almost the same way.

Newcomer tip: run defense EPA EPA, expected points added, measures how much a play changed a team's chances of scoring. Jacksonville's run defense allows about −0.23 EPA per run, 3rd best of 32. A negative number for a defense means runs against it usually leave the offense worse off.

Pick under the rule: no. The model is 1.0 point from Vegas on the margin (2 needed) and 0.2 points on the total (3 needed).

Official pick:

Sun Oct 18, 12:00 PM CT

Baltimore Ravens at Cleveland Browns

Over/under voided Oct 15

Preview check (not an official pick)

Baltimore wins by 6 or more

Vegas says Baltimore by 5.5, 42.5 total points. Our model says Baltimore by 9.7, 45.6 total points.

Baltimore should win comfortably. Lamar Jackson is back at the top of the depth chart and off the injury report, and the Ravens' offense ranks 6th in EPA per play while Cleveland's ranks 25th. Our model gives Baltimore about a 77% chance and has them winning by more than Vegas does.

Tuesday's official card was built with Tyler Huntley, because Jackson was listed Out on the week-5 report. That card had Baltimore by 4.1 and no pick on the margin. This preview uses Jackson: the schedule, ESPN's depth chart and ESPN's injury report now agree on him. Deshaun Watson starts for Cleveland.

Newcomer tip: early-down EPA First and second down set up everything else. Baltimore gains about +0.21 expected points per early-down play, 1st in the league. Cleveland's offense is at about −0.04, 20th.

The locked over/under (over 40.5) was voided Oct 15 and is not a pick. The margin check is a preview, not an official pick: the model has Baltimore 4.2 points better than Wednesday's Vegas line (2 needed). The official card has no spread pick.

Official pick:

Sun Oct 18, 12:00 PM CT

Carolina Panthers at Philadelphia Eagles

Over/under voided Oct 15

Preview check (not an official pick)

Philadelphia wins, or loses by 1 at most

Vegas says Carolina by 1.5, 39.5 total points. Our model says Philadelphia by 5.3, 46.6 total points.

Our model says Philadelphia, about a 66% chance at home, with Jalen Hurts against Bryce Young. Vegas disagrees, and has moved toward Carolina since Tuesday morning, when it had the Eagles by 1.5. Five games of Eagles offense are the reason to doubt us: Philadelphia ranks 29th in EPA per play and last in success rate, while Carolina's offense ranks 8th.

Newcomer tip: success rate Philadelphia's offense stays on schedule on only about 36% of plays, 32nd of 32. A team that falls behind schedule that often ends up punting a lot.

The locked over/under (over 41.5) was voided Oct 15 and is not a pick. On the margin the model is 6.8 points from Vegas. The official spread pick stands.

Official pick:

Sun Oct 18, 12:00 PM CT

Chicago Bears at Atlanta Falcons

Over/under voided Oct 15

Too close to call

Vegas says Chicago by 1.5, 51.5 total points. Our model says Chicago by 1.3, 45.6 total points.

Who wins is close to a coin flip: our model gives Chicago about a 54% chance on the road, and Vegas has the Bears by 1.5. The disagreement is the score. Vegas moved the total up from 49.5 on Tuesday to 51.5, while our model expects about 46, helped by an Atlanta defense that ranks 3rd at keeping offenses off schedule.

Caleb Williams is Questionable (hamstring) on ESPN's injury report. The schedule lists Tyson Bagent, and the depth chart has Williams at QB1. We use Williams. If Bagent starts, the quarterback part of this margin changes. Michael Penix Jr. starts for Atlanta.

Newcomer tip: the total The total is Vegas's guess at both teams' points added together. At 51.5, Vegas expects a shootout. Our model's 45.6 is about a touchdown lower. That over/under was voided Oct 15 and is not a pick.

The locked over/under (under 49.5) was voided Oct 15 and is not a pick. On the margin the gap is 0.2 points, so there is no spread pick.

Official pick:

Sun Oct 18, 12:00 PM CT

New Orleans Saints at New York Giants

Preview check (not an official pick)

The Giants win, or lose by 1 at most

Vegas says New Orleans by 1.5, 45.5 total points. Our model says the Giants by 2.4, 46.0 total points.

Our model likes the Giants at home, about a 57% chance, with Jameis Winston against Tyler Shough. The Giants are 3-1-1, the Saints are 1-4, and the New Orleans defense has struggled badly. Vegas has the Saints by 1.5; on Tuesday morning it had the Giants by 1.5.

Newcomer tip: success rate allowed Opponents have succeeded on about 50% of plays against the Saints, 31st of 32. Half the time, the offense is staying on schedule.

Pick under the rule: yes, on the margin. The model is 3.9 points from Vegas (2 needed). The total gap is 0.5 points, so no.

Official pick:

Sun Oct 18, 12:00 PM CT

New York Jets at New England Patriots

Too close to call

Vegas says New England by 7.5, 43.5 total points. Our model says New England by 8.0, 41.6 total points.

New England should win at home, about a 73% chance on our model. Drake Maye's passing game ranks 9th in EPA per pass play, and the Jets, at 2-3 with Geno Smith, rate 29th of 32. Our model and Vegas are only half a point apart.

Newcomer tip: opponent-adjusted rating A team's raw stats get corrected for who it has played, so beating weak teams counts for less. After that correction, the Jets sit 29th of 32 and New England 12th.

Pick under the rule: no. The gap is 0.5 points on the margin and 1.9 on the total.

Official pick:

Sun Oct 18, 12:00 PM CT

Pittsburgh Steelers at Tampa Bay Buccaneers

Over/under voided Oct 15

No official pick

Vegas says Pittsburgh by 2.5, 39.5 total points. Our model says Tampa Bay by 2.3 with Baker Mayfield, or Pittsburgh by 1.4 with Jalon Daniels, and 44.0 total points either way.

Who wins depends on Tampa Bay's quarterback. With Mayfield, our model makes the Buccaneers slight favorites (about 57%); with Daniels, it has Pittsburgh by 1.4, close to Vegas. Tampa's offense ranks 27th in EPA per play, and Pittsburgh's defense has given up the most big plays in the league.

ESPN lists Mayfield Doubtful (thumb). The schedule lists Daniels. ESPN's depth chart and the nflverse depth chart still have Mayfield at QB1, so the model's run uses Mayfield. The sources disagree, and we are not guessing who starts. Aaron Rodgers starts for Pittsburgh.

Newcomer tip: explosive plays An explosive play is a pass of 20 or more yards or a run of 15 or more. Pittsburgh's defense allows one on about 8.6% of snaps, the most in the league.

The locked over/under (over 40.5) was voided Oct 15 and is not a pick. The margin depends on the quarterback and is not an official pick: with Mayfield the gap is 4.8 points, with Daniels it is 1.1.

Official pick:

Sun Oct 18, 12:00 PM CT

Tennessee Titans at Indianapolis Colts

Preview check (not an official pick)

Indianapolis wins by 8 or more

Vegas says Indianapolis by 7, 43.5 total points. Our model says Indianapolis by 9.3, 44.7 total points.

Indianapolis should win at home, about a 76% chance on our model. Tennessee is 0-5, has scored 58 points in five games, and rates last of 32. Daniel Jones and a Colts run game that leads the league in EPA per run face Cam Ward's Titans.

Newcomer tip: rushing EPA Most runs lose a little expected value, so 26 of 32 teams are below zero per run. Indianapolis is above it, at about +0.10 per run, 1st of 32.

Pick under the rule: yes, on the margin. The model is 2.3 points past Vegas's 7 (2 needed). The total gap is 1.2, so no.

Official pick:

Sun Oct 18, 3:05 PM CT

Arizona Cardinals at Los Angeles Rams

Over/under voided Oct 15

Too close to call

Vegas says the Rams by 12.5, 51.5 total points. Our model says the Rams by 10.6, 49.3 total points.

The Rams should win easily, about a 79% chance on our model. They rate 2nd of 32, and Matthew Stafford faces an Arizona pass defense that has been one of the league's worst. Arizona is 1-4 with Jacoby Brissett and has allowed 162 points, the most of any team.

Newcomer tip: EPA per pass allowed Arizona gives up about +0.36 expected points every time the other team passes, 31st of 32. For a defense, a positive number is bad: the offense is usually moving toward points.

The locked over/under (under 52.5) was voided Oct 15 and is not a pick. On Wednesday's line the model is 1.9 points short of Vegas on the margin (2 needed). The official spread pick from Tuesday still stands.

Official pick:

Sun Oct 18, 3:25 PM CT

Los Angeles Chargers at Kansas City Chiefs

Too close to call

Vegas says Kansas City by 8.5, 42.5 total points. Our model says Kansas City by 8.7, 44.1 total points.

Kansas City should win, about a 75% chance on our model. The Chiefs are 4-0 with Patrick Mahomes. The Chargers are 0-5, and Justin Herbert's offense ranks last in EPA per play. Our model and Vegas agree almost exactly.

Newcomer tip: pass rate over expected Given the down, distance and score, a typical team passes a certain share of the time. The Chargers pass about 10.8 percentage points less than that, 31st of 32, leaning on a run game that ranks 28th.

Pick under the rule: no. The gap is 0.2 points on the margin and 1.6 on the total.

Official pick:

Sun Oct 18, 3:25 PM CT

Buffalo Bills at Las Vegas Raiders

Preview check (not an official pick)

Buffalo wins by 6 or more

Vegas says Buffalo by 5.5, 53.5 total points. Our model says Buffalo by 7.8, 52.1 total points.

Buffalo should win, about a 72% chance on our model. Josh Allen's offense is 1st in the league in EPA per play, and the Raiders, with Kirk Cousins, rate 25th of 32. Buffalo's own defense ranks 30th, which is why both Vegas and the model expect a lot of points.

Newcomer tip: explosive play rate Buffalo gets a 20-yard pass or a 15-yard run on about 11% of snaps, roughly one play in nine, 1st in the league.

Pick under the rule: yes, on the margin, barely. The model is 2.3 points past Vegas's 5.5 (2 needed). The total gap is 1.4, so no.

Official pick:

Sun Oct 18, 7:20 PM CT

Dallas Cowboys at Green Bay Packers

Preview check (not an official pick)

Fewer than 53.5 points

Vegas says Green Bay by 1.5, 53.5 total points. Our model says Green Bay by 3.4, 50.0 total points.

Green Bay should win at home, about a 60% chance on our model, with Jordan Love against Dak Prescott. Dallas has one of the league's best offenses (4th in EPA per play) and its worst defense (32nd), and Cowboys games have averaged about 55 combined points. Our model expects plenty of scoring here, just not as much as Vegas.

Newcomer tip: defense EPA allowed Dallas allows about +0.18 expected points per play, last of 32. For a defense, positive means the other team is usually moving toward points.

Pick under the rule: yes, on the total. The model is 3.5 points below the 53.5 total (3 needed). On the margin the gap is 1.9 points, just short of 2.

Official pick:

Mon Oct 19, 7:15 PM CT

Washington Commanders at San Francisco 49ers

Preview check (not an official pick)

San Francisco wins by 8 or more

Vegas says San Francisco by 7.5, 50.5 total points. Our model says San Francisco by 10.2, 50.2 total points.

San Francisco should win, about a 78% chance on our model. The 49ers are 5-0 and rate 1st of 32, with Brock Purdy's passing game leading the league in EPA per pass play. Washington, 1-3-1 with Jayden Daniels, rates 30th.

Newcomer tip: success rate San Francisco's offense stays on schedule on about 52% of plays, 1st of 32. That steadiness is a big part of a 5-0 start.

Pick under the rule: yes, on the margin. The model is 2.7 points past Vegas's 7.5 (2 needed). The two totals are 0.3 apart.

Official pick: