Arizona beat the Rams in Los Angeles as a two percent underdog, and our model lifted the Cardinals from five wins to six
Updated
Arizona went to Los Angeles as the model's biggest underdog of the season, a two percent chance, and won 18-16. A week earlier it had lost 45-21 at home to Detroit. The Cardinals are 2-4, and when we ran the rest of the season they came out at 6-11 with a two percent playoff chance, up from 4.8 wins two weeks ago and from 4.4 in July. A win and a half in a fortnight is the largest upward move any bottom-ten team has made this season, and it came from one Sunday against a roster the model had rated as the best in football.
The offense has been competent all year and the model has started to give it credit. Jacoby Brissett is 156 of 234 for 1,328 yards with nine touchdowns and six interceptions, an 83 rating, and projects for 4,351 yards and 26 touchdowns behind a 91 rating, up from 24 and 89 two weeks ago. Michael Wilson projects for 124 catches and 1,424 yards, up from 112 and 1,332, and is now the eighth-highest receiving projection in the league. Trey McBride projects for 102 catches, the most of any tight end. Kendrick Bourne has climbed from 50 catches to 65. Marvin Harrison Jr. is no longer on the roster, having been traded to Houston. Jeremiyah Love projects for 552 rushing yards and Tyler Allgeier for 298. The unit projects for 350 points, 20th in the league, up from 323 and 24th.
The defense is still the problem and still graded that way. Arizona has allowed 178 points in six games, the most in the league along with New Orleans, and our model projects the unit giving up 443 points, 31st. The Rams game was the exception that makes the number look harsh: 16 points allowed to an offense the model had at 454 projected points. Budda Baker's sack projection has nearly doubled to over four, which is the sign of a safety being used as a blitzer. Roy Lopez's sack line has doubled to nearly seven from the interior. Josh Sweat has slipped from 12 projected sacks to under eight and a half and Zaven Collins from nine and a half to seven and a half, which is the edge pair cooling off. The team projects for 41 sacks, 15th, and nine interceptions, 29th.
The division is the 49ers and then three teams fighting over second. San Francisco is 6-0 and projects for twelve wins. The Rams are 2-4 and project for ten, with Arizona now holding the tiebreaker. Seattle is 3-3 and projects for seven. When we ran the rest of the season the Cardinals finished third in the NFC West more often than fourth, which is the first run all year in which that was true, and the Rams rematch in Week 10 is the game that decides it.
The next month is a chance to bank wins. Denver at home at 58 percent, a trip to Dallas at 22, a trip to Seattle at 42, and the Rams at home at 18. When we ran that stretch Arizona went 1-3 more often than 2-2, and the Denver game is the first one this season the model has the Cardinals favored in. A 3-7 record entering the second week of November would put six wins on schedule.
The Rams game deserves the detail. Arizona went to Los Angeles with a defense that had given up 45 to Detroit six days earlier and held the Rams to 16 points, and Brissett did enough on offense to win 18-16. The model had given the Cardinals a two percent chance, which is the lowest number it has assigned to any team in any game this season, and it is the single largest upset of the year by that measure. The week before was the opposite: Detroit scored 45 in Glendale and the game was never close. Those two games are the Arizona season in miniature, a defense that can be shredded by a good offense and can occasionally hold a bad day from a good one. The Harrison trade changed the offense less than the model expected. Wilson and McBride have absorbed the targets, Bourne has climbed, and the points-for projection rose 27 in two weeks. What moves the number from here is the Denver game, the first the model has had Arizona favored in all season. Win it and the Cardinals are 3-4 with the Rams tiebreaker in hand and a six-win projection that looks low. Lose it and the model will treat the Rams game as the outlier it probably was.
The matchup record is a team that keeps losing the games it should and winning the ones it should not. We gave Arizona 20 percent against the Chargers and the Cardinals won. We gave them 28 against Seattle, 28 in San Francisco, 48 in New York and 16 against Detroit, and they lost all four. We gave them two percent in Los Angeles and they won. Two wins from about one and a half expected is slightly ahead of the projection, and the Rams game is the single largest upset of the season by the model's numbers. The market opened Arizona at six and a half wins. Our model sat two under it in July and sits a half-win under it now. The lean is about a defense that has allowed 178 points in six games. The Rams game is the reason it is only half a win.
