Game Over Report: Tracking the Likelihood of NFL Coaches Getting Fired

By Stephen Juza

June 28th, 2023

The coaching hot seat is a perennial topic in the NFL, with no coach being immune to the pressure and speculation. With so much turnover, virtually no coach will escape the dreaded hot seat conversation. Even Bill Belichick, who is within striking distance of the career win record, has slowly had the calls for his job ramp up after missing the playoffs in two of the last three seasons.

With so much pressure on coaches, the constant chatter of ‘are they on the Hot Seat’ weighs with every win or loss. A particularly bad loss is immediately followed by “Is this the game that gets them fired?” A key winning streak to close the season can give a team hope that just maybe the coach is living up to the promise and deserves another season at the helm.

This season, we will debut the “Game Over Report” where each week, we will track each coach’s likelihood to be fired by the end of the season. Using a machine learning algorithm, we will assign a probability to each coach that they will be fired by the end of the season.

Who is likely to be fired?

Since 2010, an average of seven coaches have been fired each season. It’s such a common occurrence that a frequent prop bet on various online bookies is for how many coaches will be fired, which coach will be fired, and if they will be fired by the end of the season.

So what does the average fired coach look like? A coach who just led a team to a five-win (or worse) season, who was a career 40%-win percentage coach, and ended the season in the bottom third of the league in points scored and points allowed. If coaches can avoid a terrible season, quiet mediocrity is enough to skate by. Nine-win seasons are typically enough to stick around for another season, if they play nicely with the ownership group.

Which coaches have surprised the model?

While the machine learning algorithm used in the "Game Over Report" boasts a 85% accuracy rate, there have been instances where it failed to predict coaching moves accurately.

Gus Bradley, Jacksonville Jaguars – 2013 (68% chance to be fired)

Gus Bradley’s first season in Jacksonville did not inspire much hope for the future. Finishing with a 4-12 record with the worst offense and fourth-worst defense in the league (by points scored). While he survived his initial season, he eventually was fired before the conclusion of his fourth season after leading the team to a 14-48 record.

Hue Jackson, Cleveland Browns – 2017 (67% chance to be fired)

Hue Jackson’s lack of firing surprised many people in 2017. Not only was the team coming off an 0-16 season, the team only went 1-15 in his first season with the franchise. Despite the 1-31 record, Jackson came back for a third season, only to be fired eight games into the 2018 season.

Jim Harbaugh, San Francisco 49ers – 2014 (0% chance to be fired)

While it was initially reported that the decision to leave the 49ers was a ‘mutual’ decision, Jim Harbaugh disagreed with this statement. The model had 74 instances where a coach had 0% chance of being fired at the end of the season – Harbaugh is the only one that was fired. It’s easy to see why the model was surprised. Winning 70% of his games with the franchise, one NFC championship, and playoffs in three of his four seasons. Success that most franchises would strive for, but friction with ownership led to an earlier departure than expected from the NFL.

As the NFL season unfolds, the "Game Over Report" will provide insights into the likelihood of coaches being fired by the end of the season. While no coach can truly escape the hot seat, the report will offer a glimpse into the unpredictable nature of coaching decisions in the league and keep fans and pundits engaged in the ongoing discussion of job security in the NFL coaching ranks.