Dallas won a shootout in Houston and our model moved the Cowboys from nine wins to ten, with the playoff odds nearly doubled
Updated
Dallas went to Houston as a coin flip and won 34-30, scoring 34 against a defense our model ranks ninth, and the Cowboys are 2-2 with the fifth-best offense in football by projection. When we ran the rest of the season Dallas came out at 10-7 with a 69.5 percent playoff chance, up from 8.8 wins and 37.5 percent after Week 3, and the jump is one of the two largest upward moves in the NFC this week. The model had the Cowboys at ten and a half wins in July, cut them to nine after a 1-2 start, and has now put most of that back. Sunday is part of the reason. The larger part is that the defense, which has been the problem all year, graded better in this run than in any since July.
The offense has been the best in the division since the opener and the numbers have kept climbing. Dak Prescott is 105 of 150 for 1,066 yards with eight touchdowns and one interception, a 105 rating through four games, and our model projects him for 4,531 yards and 32 touchdowns against seven picks. CeeDee Lamb has 37 catches for 498 yards and four touchdowns, the most receiving yards in the league, and projects for 115 catches, 1,565 yards and ten scores, all up from a week ago. George Pickens has 225 yards and projects for 74 catches. Jake Ferguson has three touchdowns on twelve catches and projects for eight. Javonte Williams has 231 rushing yards and five touchdowns and projects for 888 and eleven. The unit projects for 428 points, fifth in the league, and Dallas has scored 122 in four games.
The defense is where the model has changed its mind, and the change is modest. Dallas projects to allow 368 points, 21st in the league, down from 394 and 24th a week ago, and the 18th-fewest yards. Rashan Gary's sack projection has jumped from eight to twelve as his pressure rate has climbed. Donovan Ezeiruaku projects for seven. Quinnen Williams for four and a half from the interior. Caleb Downs' tackle projection has gone from 19 to 112 as the safety has taken a full-time role, and he projects for over two interceptions. The team projects for 34 sacks, 29th, and eight interceptions, 30th, which are still the numbers of a defense that does not create much. It has given up 112 points in four games. It gave up 30 to Houston and the offense covered it. That has been the formula, and it is a formula the model now rates at ten wins.
The division is open in a way it was not in July. Philadelphia is 2-2 and projects for nine wins after losing at home to the Rams. The Giants are 3-1 and project for eight, with the head-to-head already won against Dallas in Week 1. Washington is 1-3 and projects for four and a half. When we ran the rest of the season the Cowboys won the NFC East more often than any other team, which is a change from a week ago when the Eagles held that position, and the Philadelphia game in Week 7 is where the model expects it to be decided.
The next month is the most favorable stretch in the division. Tampa Bay at home at 76 percent, a trip to Green Bay at 66, a trip to Philadelphia at 60, Arizona at home at 74, and a trip to Indianapolis at 64. When we ran that stretch Dallas went 4-1 more often than anything else, and the model has the Cowboys favored in all five, which it does for very few teams. A 6-3 record entering November would push the projection toward eleven.
The player lines are an offense holding its grade and a defense finally earning one. Prescott's projection dropped from 4,553 yards to 4,531 and from 33 touchdowns to 32, with his interception line down from eight to seven and his rating projection up to 104.9. Lamb's climbed from 100 catches to 115 and from 1,431 yards to 1,565, one of the two biggest receiving moves in the league this week, with his touchdown line up from nine to ten. Pickens' dropped from 80 catches to 74. Ferguson's held at eight touchdowns. Williams' rushing projection fell from 920 to 888 while his touchdown line rose from nine to eleven. On defense, Gary's sack projection went from 8.2 to 12.1, the biggest move on the roster, Quinnen Williams' from 3.4 to 4.6, and Downs' tackle projection from 19 to 112 with his interception line from zero to 2.3. Hooker's tackle projection rose from 77 to 84. The number to watch is the points allowed. Dallas projects for 368, down from 394 a week ago, and the ten-win projection is built on that 26-point improvement holding. Tampa Bay at home is the kind of game that tests whether it does.
The matchup record is a team that has lost the games it was supposed to win and won the ones it was not. We gave Dallas 80 percent against the Giants and the Cowboys lost by eight. We gave them 78 against Washington and they won by 17. We gave them 34 against Baltimore and they lost by three. We gave them 50 in Houston and they won by four. Two wins from about two and a half expected is slightly under, and the Giants game is the one the model still holds against them. The market opened Dallas at nine and a half wins. Our model sat a win over it in July, under it after Week 3, and a half-win over it now. The lean is the quarterback and the receiver, who are playing at the level the model expected in the summer. The defense has started to be less of a liability, and that is enough to make the difference between nine wins and ten.
