More Thoughts on Firing the Coach

In the Boulder Daily Camera today, @BrianHowell33 has an in-depth look at college football programs who have struggled under the first two years of a coaching regime. The topic is of course on people's mind here in Boulder because the Colorado Buffaloes football team is suffering through its worst season ever and 7 seasons without a winning record. So a question faced by the University administration is what, if anything, to do about it?

Howell writes:

Today, Colorado (1-10, 1-7 Pac-12) will put an end to a miserable season when it hosts Utah (4-7, 2-6) at Folsom Field. If the Buffs lose, they'll finish with the worst record in school history. They are 0-5 at Folsom Field and face the prospect of losing every home game for the first time since 1891 -- CU's second season of football and a year that included just two home games. The current Buffs have been pounded week after week for the past two months.

Some CU fans are hopeful that second-year head coach Jon Embree, who is 4-20 since taking over in 2011, will be cleaning out his office before the two-year anniversary of his hire date, on Dec. 6.

Last week I discussed an academic paper that looked at the fortunes of college football programs that fired their coaches (ironically enough, with a lead author from Colorado). That paper -- Adler et al. --  suggested that programs that replace their coaches due to poor performance in general do not see improvements when a new coach is hired. Adler et al. conclude:

[A]t least with respect to on-field performance, coach replacement can be expected to be, at best, a break-even antidote.

Data provided by Howell supports the conclusions of Adler et al. but also suggests a slightly different take:

Since the 2000 season, 26 other [than Colorado's Embree] FBS coaches have won five games or less during their first two years on the job. Another 32 won fewer than nine games.

Among those 58 coaches, just eight were fired at the end of their second year. In most cases, schools gave their coaches several years to try to get it right.

That group of coaches includes several that never did -- or have yet to -- get their teams on a winning track. Duke's Ted Roof went 6-45 in four-plus seasons. Greg Robinson went 10-37 in four years at Syracuse. Stanford's Buddy Teevens was 10-23 in three years, Mississippi's Ed Orgeron was 10-25 in three years and Washington's Tyrone Willingham went 11-37 in four years.

That group also includes 18 coaches that eventually led their team to at least one bowl game (San Jose State's Mike MacIntyre will join that list this year).

Tommy West of Memphis (five bowls), Kentucky's Rich Brooks (four bowls), Kansas' Mark Mangino (four bowls), Arizona's Mike Stoops (three bowls), Ron Zook of Illinois (three bowls) and Ball State's Brady Hoke (two bowls) all led their teams to success after a poor first two seasons on the job.

Those 58 coaches represent just about half of all coaches replaced since 2000. If the other half won 9 games or more in their first two seasons, then we can conclude that "success" is indeed just about a break even proposition.

That said, the success stories are worth noting:

Of the 58 coaches since 2000 to win eight or fewer games through two years, just 18 (31 percent) of them eventually took their team to a bowl game. That includes Embree's predecessor, Dan Hawkins, whose tenure at CU would hardly be considered a success.

Of the 18, just 10 went to multiple bowl games, so clearly the odds are against Embree.

The good news for Embree and the Buffs, however, is that it can be done and when it is done, the taste of success is so sweet.

Within four years of Sports Illustrated profiling its retched program, Rutgers played in two bowl games, vaulted into the top 10 of the national rankings and gained a major victory in recruiting by beating traditional power Ohio State for tackle Anthony Davis, one of the top prep players in New Jersey. Davis was an All-American at Rutgers and became the No. 11 pick in the 2010 NFL Draft.

Vanderbilt's Bobby Johnson (4-20 through two years) led the Commodores to a bowl game in his seventh season. It was Vandy's first bowl appearance in 26 years.

So based on history, Colorado's Sean Embree has about a 17% chance of taking Colorado to multiple bowl games from this point in his career. At the same time, the record of coaching replacement presented by Howell gives about a 50% chance of 9 or more wins over the next 2 seasons, or about double what the Buffs have seen the past two seasons. Of course, given that the Buffs on-field performance can only improve, whatever decision that Colorado administrators make about the coaching staff, it will probably look like a good one for the next year or two.

More generally, are the experiences of Rutgers, Memphis, Kentucky, Kansas, etc. simply a consequence of the statistics of hindsight? Or is there actually an art to turning around football programs? If so, then Colorado might apply a bit more statistical insight into its coaching decision. The conclusion they should reach is not that the head coach doesn't matter (an implication of Adler et al.) but instead, picking the right coach makes all the difference.