As I have mentioned previously, the model's confidence levels begin to arise generally after week four or five, and then it is a sprint to the end of the season. To arrive at week four being 61% against the spread is very encouraging. 

This week we are seeing the model start to assert itself. Specifically, you will begin to see it predict dogs to win outright and larger margins of victory. Both have happened a little early this season. The model has Detroit beating Chicago as an underdog. It also predicted that Washington will lay the wood to Atlanta, winning by 14.38 points.

This will still be a light week as I prefer to wait for more data before moving on its predictions full bore. 

The model will offer two survivor pool picks for your consideration, one favorite and one dog. This week it is Dallas and Baltimore. 

Jacksonville Jaguars v Cincinnati Bengals (-7.5, 45.5)

The model has the Bengals covering relatively easily with a near eleven-point victory. It also has the game staying under by a hair. 

Still, both teams are subject to too much variance for the model to be confident to lay the 7.5. 

This will be the first time we see Meyer coach on a short week. Every coach has talked about the learning curve involved in doing that effectively. Meyer has already seemed out of his depth in the NFL. In addition to the math, the situation bodes well for the Bengals. 

The model is not taking a position in this game. 

Washington v Atlanta Falcons (1.5, 48)

The model is undaunted by Washington's collapse in Buffalo. We have heard about the Washington elite defensive line all season, yet we have not seen it. If ever there was a game to get right for the defensive line, this is it. Atlanta's offensive line ranks at or near the bottom of the league in every relevant metric. 

This will also be the easiest defense Washinton has faced all season. The model ranks the Falcons defense as the second-worst defense in the league. Washington's prior three opponents currently rank first, eighth, and fifteenth. 

The model is taking Washington -1.5; and -12.5 +400 (quarter unit)

Detroit Lions v Chicago Bears (-3, 42.5)

I wrote about this game earlier this week, so this will be brief. 

It is rare that I would not back my model when it picks a dog as an outright winner. Still, given everything that has happened, if I were to take a position on this game, I would move on to the Bears -2.5 at Pinnacle. The line move alone is worth 15% to the win and 15% to the push or win at Pinnacle at -115. 

I am not going to touch this game on either the spread or total. 

I am going to take a player prop on Detroit WR Quintez Cephus (Under 2.5 Receptions, +135)

The model is not taking a position in this game. 

Tennessee Titans v New York Jets (7, 44.5)

This is when we should start to see Zach Wilson improve. Not enough to take a side either way. The model is content to stay away from inconsistent teams, such as the Titans. The model has no problem with good teams. It has no problem with bad teams. It just does not like wildly inconsistent teams. 

If you want to have a side on this game, the model would suggest the under. 

The model is not taking a position in this game. 

Cleveland Browns v Minnesota Vikings (2, 52.5)

The line has not moved much in the last week. This is typically a place where I would love to bet on the Vikings. The model is much higher on Minnesota than most people I know. Minnesota also possesses one of the few legitimate home-field advantages in the NFL. Still, the model is going chalk. 

The ticket sales are crazy. The Browns have 67% of the tickets and 80% of the money. This is a no-brainer play that you live with easily if it goes sideways.  

The model is high on Minnesota, and especially Cousins. It backs the Vikings in almost any game against a middling team. The Vikings gave Arizona all they could handle. They lost to Cincinnati. And they beat the Seahawks. The Browns are not middling. 

The model sets the correct line at Browns -4.5. Getting them at -2 and covering the three is an automatic play. 

The model also strongly favors the over, but confidence is low. I would prefer to bet team totals over in the first half than the game. 

The model is taking the Browns -2

Indianapolis Colts v Miami Dolphins (-2, 42.5)

This is the second of three AFC South v AFC East matchups, and the model is as interested as it was in the Titans against the Jets, which is to say, not at all. 

The simulations are almost perfectly in line with the vegas numbers. This is an easy pass. 

The model is not taking a position in this game. 

Carolina Panthers v Dallas Cowboys (-4.5, 50.5)

There is a lot of specious advice flying around about this game, even from some trusted sources.  Here is the primary argument: Carolina has a ton of rest, and Dallas is on short rest. This line is a reaction to how completely Dallas destroyed Philadelphia. Dallas has always been an overvalued team, and this line being across three and four is another example of that overvaluation. Some of this is generally true, but the details are nuanced, and here this argument is misleading bettors. 

Let us start with the most crucial point - this line is a reaction of Dallas destroying Philadelphia and Dallas being a public team. If that is true, when I look up the "look-ahead" line from last week (before Dallas played Philadelphia), I better see a smaller line, or else this is simply a lie. Well, last week, on September 21, 2021, the line for this game was Dallas -5.5. 

So, the line has moved down since Dallas annihilated the Eagles. So the main argument is demonstrably false. The rest issue is real, but players will routinely state that rest is significantly less important early in the season. 

Here is what matters most if you are considering taking a side in this matchup. Dallas' defense is outperforming expectations, and it will not last. They beat the Chargers with interceptions. The Chargers moved the ball easily, as did the Bucs. Do not take anything from the performance of Dallas' defense against the Eagles' offense on Monday. The Eagles' are the most overrated offense in the league - take away the game against Atlanta and look at their numbers. It is horrible.  

The Cowboy's offense is a legitimate top-five offense in the league. Still, Gallup is out, Cooper is hurting, and Collins is out. You also never know what Dallas' game plan will be. This season, Moore has done a great job of taking what the defense is giving and allowing Dak to check out of plays at the line. 

The Panthers' defense ranks as the third best defense according to the model. They are missing Jaycee Horn at the corner now. More importantly, the Panthers played the Jets, Saints, and Jaguars. That is Zach Wilson in his first NFL start, Winston, and Davis Mills in his first NFL start. They will regress. Still, the model believes in Rhule, and he is getting free rushers. 

On offense, as I detailed recently, Darnold has played significantly better than he has ever played in his life. Yet, he has done so against only one good defense, the Saints. New Orleans had been on the road for a month because of the hurricane. Despite those advantages, the model ranks Carolina as only the 21st best offense. Darnold will also be without McCaffery. 

THE MODEL SPEAKS:
2021 NFL Week Four Projections and Analysis