With the first head-coaching casualty of the 2017 season occurring earlier this week, as the New York Giants fired Ben McAdoo, it's a perfect time to revisit our coaching hot seat model.
Back in August, we estimated the likelihood each head coach could survive a poor 2017 season using a machine learning model based on multiple data points that you can read more about here.
This time around, I'm factoring Football Power Index's current projections of each team's 2017 record into the hot seat model, estimating each coach's chance of being fired based on that projected final showing.
Although the model is highly accurate in aggregate, in individual cases, it's best to think of the probabilities as measuring how unusual or expected it would be for a coach with a similar résumé to be fired, rather than as a definitive prediction. Each situation is unique, and there are factors no quantitative model can consider.
In McAdoo's case, there's no input variable for holding uninspiring news conferences or having heated blowups with respected veterans. In fact, our model gave McAdoo just a 33 percent chance of being fired based on his résumé alone. Still, that is an unusually high number for a second-year coach, especially one a year removed from an 11-5 season that included a playoff appearance. It's that fall from grace that primarily drove McAdoo's firing probability within our model: A coach with lower expectations would have had a better chance of surviving to his third season.
With McAdoo off the board, let's take a look at the current projections heading into Week 14 of the 2017 regular season. Whose seats are hottest?
Chuck Pagano, Indianapolis Colts
Projected record: 4-12
Firing probability: 74 percent