Each week I will publish the model’s rankings in the following categories: Projected Final Standings, Overall (Power), Quarterback, Offense, and Defense.
Predictive analytics in the NFL involves a constant battle with the sample size. Week two predictions will be heavily influenced by the team’s opponent and performance in the season’s first week. The numbers will stabilize as we approach weeks four through eight, which will reduce variance.
Still, there is important information we can obtain from the first few weeks. If a team is performing significantly above or below the mean in multiple categories, we can begin to chart trends and spot weaknesses in the betting market.
Before we get to the rankings, I must confess my never-ending love of rankings. I will rank anything. I do not take most of them too seriously. I just enjoy it. I can give you the ranking of my eight celebrity dinner guests – a ranking I have honed for 30 years. I have placed Shakespeare’s top 16 plays into a March Madness-style bracket and had vigorous debates about the winner. I have a model that examines movies and suggests the most profitable plot, actors, director, and release date to make the biggest blockbuster ever. I even have a ranking for socks. Most of those rankings are wildly subjective. What follows is not. It is an entirely objective model.
So with that, let’s start ranking!
Projected Final Standings:
The data that follows is based on more than just last week’s results. It is based on countless simulations. Those simulations now include the actual results from the previous week and the model’s prior simulations. As the season progresses, the simulations become less important until they are removed entirely. At that point, earlier weeks in the season become less critical.
These rankings will change dramatically over the next three weeks. As of today, this is where the model stands. As you read through the list that follows, notice the absolute disparity in talent the model sees between the teams throughout the season. The model has clear favorites. Then compare that disparity to the rankings that follow where, after one week, it sees teams generally bunched in the middle. In short, that is the variance I have been writing about over the last week.
|AFC Teams||Projected Wins||Make Playoffs||Win Division||Win Conference||Win Super Bowl|
|Kansas City Chiefs||13||83.00%||62.76%||39.39%||24.72%|
|San Diego Chargers||10||61.00%||29.06%||8.45%||6.60%|
|New England Patriots||9||47.00%||15.14%||2.29%||1.08%|
|Las Vegas Raiders||8||43.00%||12.12%||1.47%||0.18%|
|New York Jets||5||8.00%||0.18%||0.00%||0.00%|
In the NFC, the most shocking development is the model now predicts that New Orleans is the only team that can keep all four teams from the NFC West making the playoffs. I cannot qualify Green Bay as a shocking result. It was only one game. Winston’s performance, on the other hand, was the absolute ceiling of his range of performances. So even if he regresses, New Orleans may be more talented than last season.
|NFC Teams||Projected Wins||Make Playoffs||Win Division||Win Conference||Win Super Bowl|
|Tampa Bay Buccaneers||12||82.00%||60.89%||37.07%||22.57%|
|Los Angeles Rams||12||73.00%||39.00%||15.21%||10.44%|
|New Orleans Saints||11||65.00%||34.06%||11.60%||7.54%|
|San Francisco 49ers||10||65.00%||30.00%||9.00%||5.56%|
|Green Bay Packers||10||52.00%||19.50%||3.80%||1.98%|
|New York Giants||5||23.00%||2.54%||0.06%||0.01%|
You can use the power rankings to produce a raw expected spread (RES) on a neutral field by subtracting each team’s number. The RES does not account for the scheme employed by each team. The model’s ultimate prediction may differ.
Moreover, this is the raw data from the model. I told the data to assume that last week’s games were the only NFL games ever played. I did this, despite the propensity for huge variance, to illustrate one point. Notice how the overwhelming majority of teams, 26 of them, are separated by no more than 2.21 points.
Around week four, this graph will begin to produce a highly accurate prediction on the spread in games on a neutral field. The extremes will regress to the middle. For now, it serves as a reminder that parity is more widespread than the Vegas line indicates.
The model ranks quarterbacks based on its own metric called Total Passer Rating (TPR). The metric can be faulted, like any metric can, for undervaluing the contributions of quarterbacks such as Lamar Jackson. I accept that criticism. Generally, I believe most people will not have too many issues with the ranking. You will notice that it punishes quarterbacks for interceptions. The model does not like quarterbacks that throw interceptions.
Remember this is based solely on last week’s performance. I trust Rodgers will move his way up the list, and some will inevitably fall.
Also, notice that the four quarterbacks that start for the NFC West are all rated in the top seven. The scary part is that all four may remain in the top ten at the end of the season. Based on what we saw from the NFC North, East, and South, New Orleans may be the only team that prevents the NFC West from sweeping the wildcard slots in the NFC.
|1||Matthew Stafford||14.30||After writing at length before the season that Stafford will explode this year with McVey and was the best bet for MVP, I am thrilled to see him atop the list at the end of week one. He was sensational.|
|2||Teddy Bridgewater||14.24||Prototypical Bridgewater. I do not expect him to stay at the top of this list, but I expect him to remain in the upper half. Denver will hang around all season.|
|3||Russell Wilson||13.95||Russ and Waldron were devastating on Sunday. A TD to Int ratio of 4-0. They stayed ahead of schedule, got a lead. Relied on play-action and motion. This has a chance to be Wilson’s MVP season.|
|4||Jameis Winston||13.93||The Saints got good Jameis. This is always the debate with Winston. Neither Carolina nor Atlanta looked competent. If Winston stays aggressive but under control and can avoid the monster turnovers, he has the talent to remain in the top five all season.|
|5||Patrick Mahomes||13.90||The off-balance across the body touchdown throw. Is there a player in the league more fun to watch? He is unpredictable yet consistent. He defies all explanations. My only conclusion is that Ice Man and Maverick had a secret love affair after they dispatched the Mig-28s over the Indian Ocean in Top Gun (Or maybe that is why Mav left Kelly McGinnis’ house after that sultry beach volleyball scene). Either way, Mav got pregnant. And Mahomes is their prodigy. Watch him play every time you can. He does not exist in the overwhelming percentage of universes.|
|6||Jimmy Garoppolo||12.59||Jimmy G got the ball out on time. That has always been the knock on him, just a touch too late with the ball. The 49er offense requires precision. Jimmy completed 68% of his passes for a 12+ypa. That is right on target.|
|7||Kyler Murray||12.52||Hard to believe Murray ranks seventh on this list after the performance he threw down on Sunday. He had a 4/1 TD to Int ratio. 66+% completion rate. And 9 yards per attempt. He was lights out. How man crazy-good throws can the kid make? Breathtaking.|
|8||Tyrod Taylor||12.21||Deshaun who? In all seriousness, Taylor was great. He has a history of being great at times. I cannot help but like the guy. I do not believe he will keep his spot in the top ten, but he is capable of playing this type of game.|
|9||Jalen Hurts||12.13||Temper expectations as it was against Atlanta…but wow. A completion percentage of 77, 3 touchdowns, 0 picks. 7.5 ypa. One sack. It was a phenomenal performance. In my divisional preview, I stated, “there is a universe in which the Eagles win the division.” Hurts playing like this is that universe.|
|10||Mac Jones||11.96||The best first-start rookie performance. While Miami’s defense lacks elite pass rushers, which certainly benefited Jones, who is a pocket passer, the back-end of the Miami defense is elite. Jones had a 74% completion percentage on high volume, 39 attempts. Took only one sack, no picks, and 7.2 ypa. Impressive.|
|11||Dak Prescott||11.77||Sensational performance off the gruesome injury. He played well enough to beat Brady. He thrived under pressure. Pressure rate of 40.9%, and +.32 EPA/dropback. A 3 to 1 TD/Int ratio.|
|12||Joe Burrow||11.02||I criticized them for passing on on O-Line help, but we cannot deny his connection with Chase. Burrow is part of the tremendous young QB core in the NFL.|
|13||Tom Brady||10.62||Brady excelled with pre-snap motion: 72.4% completion percentage, 3/1 TD/Int ratio, 6.72 ypa.|
|14||Baker Mayfield||10.54||Horrible pick at the end. Baker was tripped up. I hope he has going to sail that ball out of bounds. Apart from that, he looked in mid-season form. The Browns are legitimate.|
|15||Justin Herbert||9.79||It is so easy to watch him play. He is going to be great. Had for him an average day, but won on the road.|
|16||Lamar Jackson||9.76||Lamar had three fumbles, and they lost two. The effectiveness of his legs is negated with the turnovers. He threw the ball a little better than we expected. Steve Young was correct pre-game when he said Lamar has to become a much more skilled passer.|
|17||Kirk Cousins||9.28||Watching Cousins play football is like eating bland spaghetti every night for dinner. This type of performance is exactly what we expect from Cousins. High completion percentage. Decent yards. Few mistakes. Safe. Good numbers in another tough loss.|
|18||Sam Darnold||9.15||Darnold got out of the Jets game with a meh performance. Good enough to win and push the spread. He looked sharp at times. I expect him to finish much lower on this list by the time season ends.|
|19||Daniel Jones||9.01||Daniel Jones and Jason Garret are so remarkably slightly below average. If we rewarded people that consistently performed within a hair’s breadth of mediocrity, they would be on the cover of Wheaties cereal. Look for more of the same.|
|20||Derrick Carr||8.56||He has always been a little better than his reputation. He is going to spend his career between 15 and 23 on this list.|
|21||Taylor Heinicke||8.37||It is his team for the next six weeks with Fitz out. He performed above expectation in relief but still below the median. Knowing he has two months at the helm, perhaps he will move above the median.|
|22||Josh Allen||7.94||The accuracy was not there. That is the primary concern with Allen. Was last year his Wentz 2017 season? Pittsburgh has an elite defense. There are explanations. If he is off the mark next week, lookout. Buffalo can turn into the Titanic overnight.|
|23||Tua Tagovailoa||7.34||On the plus side, he won. This is the first time he has had an offense designed for him. It is akin to a first start. Hostile place and a tough defense. Big division win. He needs to be better about 12% better across the board. There is a path for him and Miami. He should be aiming for a ranking in the top 16 by the end of the season.|
|24||Carson Wentz||6.92||From Wentz, he came. He looked better than he had recently. It seems true that Reich knows how to get the best out of him. But this type of performance is precisely the QB he is. He will never come close to his 2017 numbers again. That was a fluke. Neither will he be as bad as last season. This is Wentz. A bottom 33% starter.|
|25||Ben Roethlisberger||6.64||The model is not high on Ben nor the Steelers, and one win against Buffalo that required a defensive touchdown did not change that view. It made the model seriously question Buffalo. Ben’s numbers were wholly unimpressive. 56% completion percentage, horrible. Only 5.8 yards per attempt, well below the median. Only 188 total yards, and a long of 24. I know everyone wants to jump on the Ben farewell tour. But he is a slightly below-average quarterback in the NFL. This is his neighborhood, between Wentz and Goff.|
|26||Jared Goff||6.51||Goff played decently. The system in Detroit is far better suited to him. This type of performance seems repeatable for him.|
|27||Zach Wilson||6.45||The rookie struggled. Not as bad as some rookies, but it was not a stellar performance. The fundamental aim for him is to grow and improve as he learns the pro game. If he finishes the season around 18-22 on this list, it is a massive win for the J-E-T-S, Jets, Jets, Jets!|
|28||Andy Dalton||5.78||Start your clocks on the end of Dalton and Nagy. Justin Fields is coming.|
|29||Matt Ryan||5.30||Matty, 4.7 yards per attempt? Really? Really? That is not very Matty-Ice. It is just one game. There were some drops. He is way better than his performance on Sunday.|
|30||Ryan Tannehill||4.99||I am not sure who missed who more … Did Tannehill miss Arthur Smith more, or did Smith miss Tannehill more? Tennessee’s offense looked different without Smith, and Tannehill looked worse.|
|31||Trevor Lawrence||4.72||Not a great look rook. Not great. There is a lot to build on with Lawrence. His skills were evident despite the poor outing. Do not fret, Jags.|
|32||Aaron Rodgers||3.22||There is nothing to say. It was an outlier. Rodgers will be fine. Just forget this game and move forward.|
Offensive & Defensive Rankings:
To read these graphs correctly, you need to know that a team that ranks precisely at 1.00 is considered league average. Technically it is the 16½ best team on offense or defense. Know the exact median provides a standard candle for the model to compare units. Thus, in the chart below, The Saints, on the far right, performed 44% above the league median on offense last week. The Packers, on the far left, performed 80% below the median. It is a pure coincidence that those two teams played each other.
The lowest ranking team will be around 20% below the median as the season moves forward, and the highest will be 20 to 30 percent above the median.
On the defensive side of the ball, we had even more extreme, outlier performances last week. The Saints rating indicates they performed 400 percent over the league median. That will not last. They will crash back to earth. Conversely, even some of the worst defensive performances were still within expectation as they were 33% below the league median at worst. That is not common, but 20 plus percent is, so it is within reason.