NFL Coach Firing Model: Week 8

By Kurt Bullard and Harrison Chase

Following a blowout loss to the Jets in Week 4, Joe Philbin was the first coach to be handed a pink slip by upper management this year. Philbin however, has not been the only coach to be in the hot seat this year. The Lions just fired offensive coordinator Joe Lombardi, with rumors that Jim Caldwell may be next if there isn’t a quick turnaround. Is there truth to the rumors, and is Caldwell the only one who shouldn’t get too comfortable in his home city?

Last year, we implemented a coaching firing model to assign a probability that a front office would actively let its head coach go by year’s end that is designed to answer this question. We implemented it in the preseason this year and found that Philbin had the second-highest probability of being fired, which somewhat validates our model across different years. Despite the fact that we felt good about our inputs, we worried that we might be overfitting a bit, and so we decided to use several methods of cross-validation to evaluate our model. After testing various models using leave-one-out and 10-fold cross-validation methods, we found nine variables to be explanatory of firing odds:

  • Season winning percentage
  • The difference between current season winning percentage and the winning percentage by team the year before the coach was hired
  • How often the coach has made the Divisional Round with the team
  • Strength of Schedule (per SRS from Pro Football Reference)
  • How often the coach has won the Super Bowl
  • Whether a new GM has taken over the organization
  • Separate dummy variables for whether a coach is in his first year, second year or, third year

All of the variables make intuitive sense, as they reflect absolute performance, relative performance compared to expectations, track records of success, and how much of the owner’s goodwill the coach has used up. To be clear, this model takes into account firings, not when a coach leaves on his own, like Jim Harbaugh and John Fox last season, who weren’t fired so much as they mutually agreed to part ways with the 49ers and Broncos. For more of an in depth description of the model, feel free to visit our post from earlier this year.

One of the more challenging aspects of creating an in-season model is projecting the rest of the NFL season, because obviously performance throughout the entire year affects whether or not a coach is fired. We used a Monte Carlo simulation to run through multiple seasons, utilizing SRS rankings from pro-football-reference.com as a measure of team strength. After iterating through 1,000 seasons, we then were able to assign a probability of being fired to each coach. The estimated probabilities for the 31 teams that have not already fired their coach are below.

 Screen Shot 2015-10-27 at 1.09.51 PM

Jim Caldwell should be feeling some pressure, as he has the highest probability of being fired at 64%. The Lions have underperformed this year, tumbling to a 1-6 start through a relatively easy schedule only ten months after a controversial pass interference call stymied Detroit in the NFC Divisional Round against the Cowboys.

Behind him, though, are three coaches who are early in their tenures with their team. Ken Whisenhunt has the second-highest probability in the league at 56% at the helm of the Titans, while second-year coach Lovie Smith and third-year coach Gus Bradley are just behind at 47% and 45%, respectively. While coaches usually get a slight pass in their first three years with a team, the performance of the Bucs, Jags, and Titans has been so poor that this effect is overwhelmed. The Titans in particular, who hired Whisenhunt after a mediocre 7 – 9 season only to watch him go 3 – 19 so far, must be feeling some pressure to change coaches

Also relatively high up on the list is Jason Garrett of the Cowboys. After being the favorite in 2014 to be fired first, he saved his job with a playoff appearance. After starting 2-4, he has worked his way back onto the hot seat. But, our model does not take into account injuries, which have decimated the Cowboys this year after a relatively healthy 2014. That may alleviate some of the pressure on Garrett, although with Jerry Jones as an owner you never know.

Chuck Pagano has also been under the microscope of the media, exacerbated by the Snapfu against the Patriots. Our model, however, only gives Chuck a 14% chance of being fired, buoyed by his success in the last two years.

On the other end of the spectrum, Bill Belichick has the lowest probability of being fired, with a .04% chance of this year being his last in Foxborough. After years of uncertainty about the future of the Jets head coaching positions, Todd Bowles looks like he’s safe for the year, having led the Jets to a surprising 4-2 start. And, despite the controversy surrounding Chip Kelly’s personnel decisions and the poor play of Sam Bradford and DeMarco Murray, the former Oregon coach only has a 6.8% chance of Howie Roseman and Co. deciding to part ways with the Philadelphian pariah. All in all, our model estimates that 5.28 more coaches will be fired by years end, a total that would, when adding in Miami’s firing of Philphin, be right on track for most years.

Jim Caldwell’s job is on the line, while Belichick is yet again a New England fixture. All is normal is the NFL.

About the author

harvardsports

View all posts

1 Comment

Leave a Reply

Your email address will not be published. Required fields are marked *