NFL Week 14: Progression and Regression

Last week’s introduction to my NFL model proved to be fruitful for those that followed along, going 4-1 and yielding +10.01 units. This brought the model to a 61-34-4 record that has yielded 86.45 units and a 22.94% ROI (Note: the model record in last week’s newsletter had a few miscounted games, and has since been corrected). Yes that means the one play I chose to highlight for last week’s newsletter was the only losing play, which is a bit unfortunate. But enough about that, this week I wanted to dive into another model play and highlight some key model concepts relevant to the game.

For Week 14, we will be looking at the Monday Night Football game between the 6-5-1 Minnesota Vikings and the 7-5 Seattle Seahawks. Before we dive into this game, I would actually like to rewind quite a bit to demonstrate what I believe to be one of the model’s strengths. As I was building the model this offseason, I colllected various data from the 2017 season and used hand-made adjustments for each team based on coaching changes, player movement, scheme changes, talent development trajectories, and more. The Vikings were coming off a strong 2017 campaign in which they finished one win away from a Superbowl appearance. On top of that, they added Kirk Cousins as their long-term solution at quarterback, were getting a returning Dalvin Cook, and had a plethora of young talent on both sides of the ball that were set to take the “next step” in 2018. Following the preseason, the model had the Vikings as the fourth best team, expected to win 10.40 games.

The Seahawks on the other hand were a very unimpressive 9-7 team in 2017 that performed well above their 6.57 adjusted Pythagorean expected win total. The team then proceeded to lose Richard Sherman, Michael Bennett, Paul Richardson, Sheldon Richardson, Jimmy Graham, and more in the offseason, and were set to be without their top wideout Doug Baldwin for an indeterminate amount of time. They also burned their first round pick on USC running back Rashaad Penny, which I thought was an awful use of the pick given that they already had the serviceable Chris Carson and had more glaring team needs. With all of this in mind, I’m not ashamed to share that by the end of the preseason the Seahawks were the 29th best team in the model, expected to win only 6.31 games in 2018. 

As I highlighted last week, the model is good at dynamically identifying what parts of the game teams are good at and what parts of the game being good at matters. After just a month into the season, the model had already downgraded the Vikings to a league average team (16th) and adjusted their expected wins to 8.36 The Vikings have since floated around those numbers. Seattle’s progression has been a bit more gradual as the model gave more credence to their performance week after week (29th – 28th – 27th – 24th – 23rd – 22nd – 18th – 14th – 12th – 11th – 9th – 6th – 6th). It was these two massive corrections and swings that made this game so interesting to me, because it is certainly outside of the norm for the model.

As of the conclusion of Week 13, the model has only changed 18/32 teams’ expected win total from the preseason by more than one game. That may sound like a lot at first, but a deeper dive into those eighteen teams paints a different picture. Eleven teams have lost an expected win or more over the course of the season, but seven of those teams are either currently starting (or have started for a significant portion of games) a backup quarterback and have lost expected wins in the process. On top of that, we have the Browns who ditched Hue Jackson (the model doesn’t project coaching changes) and have since crossed that one-game threshold in the positive direction. So if we take out the teams that have been subjected to backup quarterbacks or head coaching changes, the model has been within one expected win for 22 of the 32 teams. That’s not too bad if you ask me.

Back to the actual game at hand, the model has a lot to like about the Seahawks this week. Looking at the model’s five factors, the Seahawks possess a significant advantage in pass offense, rush offense (MIN has the league-worst), and special teams (shoutout to Michael Dickson). On the other side of the ball, the Vikings have a very slight edge in pass defense and a significant advantage in rush defense. Circling back to the model’s dynamic correlation coefficients discussed in last week’s newsletter, Seattle’s advantages amount to much more with offense contributing to league-wide success three times as much as total defense. As a result, the model gives the Seahawks a 56.25% probability of winning on a neutral field, and after factoring in the various game-specific factors which include Seattle’s second-best home field factor and the Vikings being in a back-to-back and three-in-four travel spot, the model makes the game SEA -6.5 which puts us in a position to lock in a SEA -3 -100 wager risking 3.514 units to win 3.5 units.

Thanks again for reading. The response and feedback I received for last week’s write-up was overwhelmingly positive and encouraging, so thank you for that. If you would like to follow the rest of the model’s plays, make sure to follow me on Twitter.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.