2021 NFL Week Two Projections and Analysis: Part One

2021 NFL Week Two Projections and Analysis- Part One

The model only has four sides as of the time of this publication. Any other sides, which may arise between now and kickoff, will be published at BLS. The model is backing the Raiders +5.5, the over in the Dolphins/Bills game, The Rams -3.5, and the 49ers -3 (+100). It is uncommon for the model to find itself backing a 3+ point favorite. It is uncommon for the model to support more favorites than underdogs. 

The first week had some performances that lend themselves to extreme reactions. The Packers, Saints, Eagles, and Falcons leap to mind. Today, one of the models’ sides is premised on the belief that the market overvalues the Eagles after they trounced the Falcons. The Rams play is as much about the model’s lack of respect for the Colts as it is about the Rams. This will be the second week in a row the model faded the Colts. That trend will likely continue as the season progresses. 

Before moving forward, let us take a look back at the first week. The model made eight wager recommendations and finished the week 7-1, +5.9 units:

TNF Over 52Won1.1 units1 unit
Seattle -2.5Won1.1 units1 unit
San Diego +1Won1.1 units1 unit
Miami + 3Won1.1 units1 unit
Denver -2.5Won1.1 units1 unit
J Williams Over 38.5Won1.61 units1 unit
Cleveland + 6Won1.1 units1 unit
Baltimore -4Lost1.1 units1.1 unit
TOTAL5.9 units

While those numbers are impressive, if you have read my articles on BLS, you probably know what I am about to say, do not read into small sample sizes. Apart from the record itself, the encouraging note from the performance above is that the model did not get “lucky.” Perhaps you can say the Dolphins outright win was lucky, but not the cover. If anything, the model was unlucky in Baltimore with the fumbles and some uncharacteristically poor game management by Harbaugh. 

The model bets half units until about week four or five. I will advise when it has accumulated enough in-season data to reach a more tolerable level of confidence. At that time, the model switches to a quarter Kelly-Criterion to determine unit size. 

A Word About the Model

No matter what the data shows in the graphs below, the only plays the model recommends are in bold at the bottom of each game’s write-up. One team may have value on the moneyline or spread, but the model backs off because it is simply not enough value when we account for variance. 

Las Vegas Raiders v Pittsburgh Steelers (-5.5, 47)

The model sees both of these teams as league average. That indicates taking the points is the strong play here. Do not overreact to the Steelers beating the Bills last week. The Steelers offense only managed 13 points against a sub-par Bills defense. (Special teams scored a touchdown for the Steelers). 

On top of that, Ben’s numbers were awful. The model ranked him as the seventh worst-performing quarterback last week. Ben had a 56% completion percentage and averaged only 5.8 yards per attempt. Both are well below the median. He threw for only 188 total yards – this is 2021, not 1973, right? His longest completion was 24 yards. He had the second-lowest ranking on passes from a clean pocket. 

The Steelers will have to figure out how to pass the ball more efficiently and with greater air yards per attempt. Conversely, they will need to switch to the game plan Philadelphia employed against the Falcons. Hurts’ average depth of target (adot) was only 3.5 yards, but he was the most accurate passer in the league last weekend.

Pittsburgh’s misguided offseason focus on improving their running game resulted in an average of only 3.6 yards per attempt. If you remove Chase Claypool’s one rush attempt for 25 yards, the Steelers run game produced 50 yards on twenty carries, 2.5 yards per attempt. First-round selection Najee Harris averaged 2.8 yards per attempt. The Steelers expected points added for every run play through one game -.33 points. On passing plays, the Steelers are a +.06 EPA. Carr and the Raiders are 2.5 times better. 

There is simply no data that supports laying the points with Pittsburgh. The only narrative for the Steelers is that the Raiders played an overtime game on Monday night and are flying east for the game.  So short rest and extended travel. I am not buying it. Lots of early Raiders’ money moved the line from -6 to – 5.5 early in the week. 

“Please understand, this does not mean the Steelers cannot cover. It simply means the math finds the range of outcomes the Steelers do cover far less likely than the range of potential outcomes they do not cover. This is not a lock or guarantee. The model deals in probabilities, not certainties.”

I do not expect Carr to throw for 400 yards against the Steelers. But I expect him to keep the game close if for no other reason than the math does not support the conclusion that Pittsburgh is capable of covering such a number without a fluke, like a special teams touchdown.

The Raiders’ defense was wildly effective against a sub-par Baltimore offensive line. They only blitzed twice the entire game yet had more pressures than any other team in the league last week. The model is still pumping the brakes a little with the Raiders’ defensive line, but the Steelers also have a sub-par offensive line.

When the line touches +6 again, the model will grab it, bringing an additional 6% advantage to the push. If it does not, I will still take the +5.5. 

For the prop bettors, the model likes Waller +115 to score a touchdown. 

This game also offers value on the moneyline. It is listed between +200 to +210. The model sees the proper line at +125

I suspect many people will tease this game across seven and ten and pair it with another play. I will confess that I do not play many teasers. I will remind you that you need each side to cover 74% of the time in a typical teaser. (A -120 line requires a win percentage of 54.55% to break even. To determine the probability of two independent results, you multiply their individual probabilities. Consequently, if both independent events are 74% to occur individually, the probability that they both occur is .74^2 = 54.77%).

Moving from +5.5 to +11.5 captures 29% more outcomes as you move across 6, 7, and 10. Those numbers make a teaser on this side palatable. 

The model will wait for a +6 to pop on this game.

New Orleans Saints v Carolina Panthers (3.5, 44.5)

Taking the Saints and laying the points will be the public’s most heavily bet side on the board this week. The Saints, not surprisingly, are on 63% of the tickets sold, but those bets represent only 26% of the handle. That means the 37% of people betting the Panthers are betting in significantly higher denominations. This indicates syndicate money is on the Panthers.

Fortunes have been lost reading tea leaves and chasing syndicate money. Most often, by the time you recognize the side, you are too late to take advantage. Moreover, syndicates lose as well. They are not impervious to being wrong. 

The model wants nothing to do with this game, and I agree with it. The model is turned off solely by the number of injuries the Saints are facing:

Of particular concern to the model is Lattimore. If not for the injuries, the model would be backing the Saints right along with the public. 

The model is aware of the big-money group’s love affair with the Panthers this season. It disagrees. Strongly. The theory is that Darnold will be so devalued that the lines will be bent toward the Panthers. This group also feels that Darnold’s struggles for the previous three years can be laid at the feet of his inept coaches with the Jets. That theory was buoyed by the Saints’ smackdown of the Packers last week. The Carolina backers see a grossly overvalued Saints team on the road against an undervalued Panthers club. They pounced.

The model sees something different. The model sees Sam Darnold and says, meh.

The Panthers won, but they were not particularly impressive last week. The Saints were otherworldly. While neither trend is likely to continue to that degree, the Saints have the better talent, especially at quarterback and head coach. Sam Darnold has yet to prove anything. If Winston does not turn the ball over, this could easily be a ten-plus point victory for the Saints. 

The model is happy to sit on the sideline for this game and just let this play out another week. Winston can be more consistent and an upper-echelon quarterback. The quarterback with the most upside in this game plays for the Saints. I will find it unendingly funny if the issue with Winston all along was that he just needed Lasik.  

The model has no bet. If I miss this train, I can live with it. 

The Model has no side in this game. 

Houston Texans v Cleveland Browns (-12.5, 48)

The model cannot justify betting on the Texans, and it will not lay the points in this scenario. Despite the tough loss in Kansas City last week, I believe it will encourage the Browns. It cemented their status as the best team in the NFC North. 

Baker, up until his last offensive play, was sensational. I still believe he was trying to sail that ball out of bounds but was tripped up. They played without OBJ, which is the last piece this team needs. He can stretch the field and be the dependable receiver Baker needs. If nothing else, through eight quarters in Arrowhead, the Browns have conclusively proven they can hang with the Chiefs. Stefanski’s coaching performance was masterful. 

Conversely, pump the breaks on the Texans are “good” argument. It was one game. It happens. If you watched Taylor and Lawrence throw the ball in the game last week, and left thinking Taylor is the more talented quarterback, then I think you are mistaking fortune for skill. Lawrence made some elite throws. Taylor’s performance owes more to the whims of good luck than to a repeatable skillset. 

The model has no side in this game. 

Denver Broncos v Jacksonville Jaguars (6, 45)

Bridgewater played fantastically against the Giants. He ended the week as the model’s second-rated quarterback. It will not last, but he will remain in the top 15. Bridgewater at his ceiling can help Denver win several games this season. The issue with Teddy is this is as good as he will ever perform. With a solid team, he will beat worse quarterbacks. His ceiling does not project as a guy that could ever beat elite quarterbacks. As of for this game, he is the better quarterback than Lawrence. 

Welcome to the league Meyer and Lawrence. Of all the harsh introductions into the league, having this Texans team take you to the woodshed has to rate among the toughest to swallow. Especially for two guys that are entirely new to the concept of losing. Gone are the pre-season, second-quarter attaboys:

Gone are the 1980’s hair-band rock-star head tosses performed before a nearly empty stadium:

The season has started. 

Despite the scoreboard and the three interceptions, there was plenty to like from Lawrence.

He made five high-level throws during the Texans game. These are elite throws that demonstrate a repeatable skillset. Conversely, Taylor only made two in that game. Lawrence retains all of the optimism we had for him when he was drafted. 

The model has no side in this game.

San Francisco 49ers v Philadelphia Eagles (3, 50)

The model is going to lay the points with Garrapolo. 

The model believes two incorrect overreactions are influencing this line. First, people are overreacting to Philadelphia’s performance against the Falcons. Hurts looked good, but his average depth of target was only 3 yards. That is a problematic approach for any offense. The Falcons defense was horrific. Philadelphia may be better than we all thought, but the model takes very little from the Falcons game. 

It is also possible that the line is overreacting to how the 49ers allowed Detroit to finish the game last week. As much as I made fun of Campbell for his kneecaps line, his team did fight for 60 minutes. There is also good reason to believe that the Lions will run the ball with some success all season long. 

The model left last weekend thinking the 49ers are a legitimate contender, and the Eagles are not as good as they looked. The model is going with the superior coach and quarterback and laying the three points. It would be off the game, at -3.5.  

The model is taking the San Francisco 49ers -3

Los Angeles Rams v Indianapolis Colts (3.5, 47.5)

The model does not have much respect for the AFC South. In particular, it does not have much respect for the Colts. Stafford should be able to gash this defense just like Wilson did. The one advantage the Colts may have is that since Seattle’s OC was in LA last season, many of the concepts will be familiar. They will either anticipate it better or relive a similar nightmare. 

I did not expect the lines to catch up this quickly in the season. I anticipated a line of -2. Moving across three stings, but I will trust the offseason work, data assimilation, and fit between quarterback and coach. 

The Model is taking the Rams -3.5. 

Buffalo Bills v Miami Dolphins (3.5, 47.5)

Both teams combined for only 33 points in their respective games last week. The total in this game opened at 49 and dropped as low as 47.5. The model is on the over in this game for several reasons. 

Neither team is facing as significant of a defensive test this week as it did last week. Miami has an excellent defense, but they are strong in the secondary, not on the defensive line. They do not have an elite pass rusher. The Steelers have an elite front line that pressured Allen constantly without blitzing. This allowed the Steelers to drop and help in coverage. The pressure combined with the tight windows resulted in a horrible game for Allen. 

His poor performance gave a megaphone to the voices that had concluded Allen would suffer a massive setback this season. The issue with Allen has always been his accuracy. No quarterback in the league’s history has ever improved their accuracy as much as Allen did last season. I am not buying the argument yet. The sample size is just too small. 

Moreover, very few teams in the NFL have the personnel to do what the Steelers did to Allen last Sunday. It was essentially the same thing the Buccaneers did to Mahomes in the Super Bowl. Mahomes looked human there as well.  

The Dolphins will not have the luxury of dropping into coverage. They will have to blitz often. The Bills will send five out and attack the weakest link. Allen will have bigger windows in which to throw and one-on-one matchups to exploit. If he is inaccurate this week against this defense, then Buffalo drops to Defcon 3. 

The Bills’ defense is not elite nor as intricate as the Patriots’ defense. Tua will have more time this week, and he will have more weapons. The Dolphins will get Will Fuller back to go with Waddle and Parker. With more weapons and against a lesser defense in week two, the model expects Miami to hang with the Bills on the scoreboard. 

The model is taking over 47.5.

+ posts