By Harrison Chase and Kurt Bullard
This past Monday, news broke from the Los Angeles Rams that devastated the masses. Despite being on the precipice of two historic records—most single-season punts inside the opponent’s 20-yard line and the all important all-time loss record—Jeff Fisher was let go by the Rams. This marked yet another ripple in a year of tumultuous change; just as the sun would rise and set, so too would Jeff Fisher go 7-9. Now, that constant is no longer.
However, the coaching carousel is only warming up, as is evidenced by breaking news that the Jaguars relieved Gus Bradley of his duties Sunday afternoon. In addition, Rex Ryan is reportedly getting a pink slip after Week 17.
A few years ago, we created a quantitative model that would predict, using a few key figures, whether or not a coach would be fired at the end of the season. Rather than speculate based on rumors in the news that can be somewhat unreliable, we thought using quantitative metrics of performance could be a more consistent and less biased way of predicting what front offices would decide once the regular season came to a close.
We have altered the model slightly throughout the years, but as of the beginning of this season, the variables we include are as follows:
- Win percentage in the given year (which we use the aforementioned power rankings to estimate)
- Win percentage in the previous season
- Three separate dummy variables for the first three years of a coach’s tenure
- How often the coach has made the divisional round
- Strength of schedule (also estimated by the power rankings)
- Super Bowls won per year
We then added a linear term for the number of years a coach has been with the team, a dummy variable for whether the team had a winning percentage above 38%, and the interaction between those two terms. We removed the dummy variable for whether the team’s GM had changed during the coach’s tenure.
At the beginning of the season, we ran the model and saw that Mike McCoy had the highest chance of being fired by season’s end. McCoy had a 55 percent chance of getting fired, The Lions’ Caldwell was assigned a probability of 39 percent, while Jeff Fisher came in third with a 38 percent chance of having his LA Story coming to a close.
But expectations have changed throughout the season due to performance. Part of the model simulates a team’s record, since winning percentage is an important variable in the model. We used NFL Power rankings at the beginning of season to measure team strength, but for this rendition we used SRS from Pro Football Reference. The Lions entered the season with a low chance of performing well in a stacked division, according to NFL.com. But, while Caldwell entered the season with a high chance of receiving a pink slip, his team now sits atop the NFC North and in second place in the conference, which has cooled his seat significantly.
The following are the odds of each team parting way with their coach as it stood entering Week 15:
Gus Bradley ranked atop the rest of the pack with a 68 percent chance of seeing his time is Jacksonville come to an end heading into Week 15. This may not come to a surprise to most, since he was 14-47 in his time with Jaguars—a winning percentage that would still be lower than Bill Belichick’s if he lost every game for 41-straight seasons. In the context of our model, he had a poor winning percentage with no playoff history, and also was out of the three-year grace period that we found organizations generally give coaches starting out.
Jeff Fisher had a 58 percent chance of getting fired by season’s end according to our model (not pictured in graph). He, too, was out of his grace period and with no playoff appearances in his tenure with the team, it was unsurprising (from a quantitative standpoint) that the Rams took action midseason. Thoughts and prayers to the King.
Marvin Lewis is also a strong candidate to be replaced. Marvin Lewis has not won a playoff game since taking over in 2003—a huge negative in our model—and is finally so far out of playoff contention that Cincinnati may part ways with Lewis who always been a good coach—just not a great one.
There’s also a pack of coaches who hover in limbo, sitting below 50 percent but still with high chances of getting fired. Chip, Fox, Hue Jackson and Todd Bowles make up the core of this group.
Most coaches, however, sit at 10 percent or below and can take a deep breath. This group includes Caldwell, who saw his odds drop 30 percent over the course of the season, and now has an 8 percent chance of leaving Detroit at season’s end.
There’s still three weeks left of the NFL season, and a lot could happen. And while we—as college seniors about to enter the real world—obviously do not wish for anyone to get fired, our model predicts on average that five coaches (including Fisher and Bradley) nonetheless will be terminated. But now, with Gus Bradley and Jeff Fisher already gone, Marvin Lewis can only glance over longingly at the cool throne from their hot seats.