2016 NFL Wild Card Predictions

We use team statistics to create models that predict outcomes of NFL games. Our models generate win probabilities and point spreads, and we consider six different models, each of which generates a unique prediction for a given game. Each model considers metrics from a number of different areas, such as turnovers, first downs, points scored, pace of play, offensive and defensive efficiency, time of possession, drive statistics, point differential, as well as a few metrics that we have created ourselves. Beyond these statistics, the models also consider factors such as “did a team make the playoffs last year?” and “how far is the away team traveling?” The article below focuses on the four Wild Card matchups for this coming weekend, but a breakdown of predictions for the entire 2016 NFL Playoffs can be found here.

Round 1 Matchups

Raiders @ Texans (-3.5)
All of our win probability models favor Houston, with win probabilities ranging from 0.56 to 0.85, while our point spread favors Oakland in a close game. Oakland is tricky, as their season-long stats likely over-estimate their current abilities (due to the Derek Carr injury). Carr has only missed a game and a half, so the Raiders still rank highly in many offensive categories used by our models, such as points scored, trips to the red zone, red zone TD rate, and winning percentage. Therefore, these predictions would favor Houston even more strongly if the data accurately reflected the quality of Oakland’s offense without Carr. Despite these issues with Oakland, Houston ranks well enough in defensive categories to favor them in all of our win probability models, as they only give up 16.6 points/game at home this season. Under Bill O’Brien, Houston has performed extremely well in their 16 games as a home favorite: 14-2 straight up, 11-4-1 against the spread (ATS), with the over going 6-10. This trend has held strong in 2016, where Houston was 6-0 as a home favorite (3-2-1 ATS), beating CHI, TEN, IND, DET, JAX, and CIN.

The table below gives our predicted spread for this matchup, as well as win probabilities for the home team from six different models. The Model 1 value of 0.85 means that this model gives Houston an 85% chance to win (and Oakland a 15% chance)

Home Away Our Spread Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
HOU OAK OAK by 2 0.85 0.85 0.76 0.56 0.65 0.79

Prediction: HOU 20, OAK 17


Lions @ Seahawks (-8)
While the public perception of Seattle may be slightly down this year, their metrics as a home team have remained impressive. They outscored their opponents by an average of 30-17 at home in the second half of the season, and went 5-0 in home games outside of their division (over the last four years, Seattle is 18-2 as a home favorite vs. non-division teams). Detroit went 3-5 on the road this season, went 0-5 against playoff teams (lost @DAL, @GB, vGB, @NYG, and @HOU), and lost their final three games to finish the season. There is also the issue of Matthew Stafford’s finger injury, which has had a hand in the Lions averaging 6 fewer points per game over weeks 14-17. Our point spread model clearly thinks Detroit will keep things close, while other models identify areas where Seattle has a distinct advantage: rushing defense, points against, playoff experience, and win quality (i.e., did you beat good teams?). This game highlights a problem that arises when considering multiple models: they don’t always agree with each other. It is always good to have multiple data points, but, when they disagree, it becomes a question of which one you trust most. In this case, Models 1 and 4 have been the most accurate in predicting previous games, and they agree that Seattle is the heavy favorite.

Home Away Our Spread Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
SEA DET SEA by 2 0.99 0.92 0.77 0.91 0.58 0.46

Prediction: SEA 24, DET 17


Dolphins @ Steelers (-10)
All of our models heavily favor the Steelers, with four of the six probabilities exceeding 0.90. The Steelers best the Dolphins in most categories being utilized by our models, including: defensive rushing efficiency, points scored, points against, game time leading, rushing first downs, trips to the red zone, and more. Pittsburgh is 11-1 as a home favorite over the last two years (7-5 ATS). While it is no surprise that the Steelers are capable of scoring points, here are a few extra reasons their offense should have success on Sunday:

  • Le’Veon Bell averaged 5.2 YPC and 139 rush yards in his last six games, while the Miami rush defense has been among the worst in the NFL, giving up 4.9 YPC and 124 rush yards/game over the same span
  • On the road, Miami has the 2nd lowest time of possession in the NFL, at 26.5 minutes/game
  • Miami’s defense ranks 21st in sack percentage, while Pittsburgh’s offense ranks 2nd in sack percentage allowed

Miami has been surprising this season, crushing their preseason Vegas win total of 7 en route to a 10-6 finish and their first playoff birth since 2008. They also finished the season by winning 7 of their final 9 games (6-3 ATS). Although they have been winning, their scoring has been extremely inconsistent: over weeks 11-17, they have scored point totals of 26, 31, 34, 34, as well as totals of 6, 14, 14 (they rank as the 6th most inconsistent scoring team in the league over that span).

Home Away Our Spread Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
PIT MIA PIT by 6 0.98 0.92 0.87 0.92 0.77 0.97

Prediction: PIT 30, MIA 21


Giants @ Packers (-4.5)
This matchup is by far the most competitive according to our win probability models, with most probabilities hovering closer to 50/50. Conversely, our point spread model has GB favored by 10 points (this model heavily weighs points scored, rushing efficiency, and how frequently a team makes it to the red zone, which are all areas where the Giants struggle). Over the past three years, the Packers are 12-2 as a home favorite outside of their division, going 8-4-2 ATS in those games. On the opposite end of the spectrum, the Giants are 2-7 (2-5-2 ATS) as a road underdog outside the division over the last three years (including 0-3 this year). Another notable trend with the Giants is how much worse they play when they are an underdog. Over the past four years, the Giants’ winning percentage is 54% higher when they are a favorite, compared to an underdog (the league average is around 21%, and only the Texans have a higher mark at 60%). In other words, the Giants have played extremely well as a favorite (they win most of the time) and extremely poorly as an underdog (they lose most of the time).

Home Away Our Spread Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
GB NYG GB by 10 0.42 0.75 0.68 0.50 0.75 0.90

Prediction: GB 23, NYG 20