AP Preseason vs. Final Polls

When it comes to picking a champion prior to a season beginning, it makes sense to stick with the prohibitive favorite. But if one is trying to pick a national champion, the AP Poll can lend a good framework with some exceptions. Take a look at the preseason Top 10 poll from the 2010 season. There were two teams, Florida and Texas, that started the season ranked in the Top 5 but that did not even finish in the Top 25. Texas finished with a 5-7 record. Nebraska, Virginia Tech, and Iowa finished outside of the Top 10, with Virginia Tech finishing 16th, Nebraska 20th, and Iowa outside of the rankings. So which teams managed to finish in the Top 10 that started in the AP Top 10? Alabama (1st to 10th), Ohio State (2nd to 5th), Boise State (3rd to 9th), TCU (6th to 2nd), and Oklahoma (7th to 6th). Two of the teams that started the season in the Top 10 finished the season in the Top 5. 

AP 2010 Preseason Top 10 Poll

1	Alabama (54)	0-0	1491
2	Ohio State (3)	0-0	1400
3	Boise State (1)	0-0	1336
4	Florida	0-0	1237
5	Texas (1)	0-0	1223
6	TCU	0-0	1160
7	Oklahoma (1)	0-0	1104
8	Nebraska	0-0	1033
9	Iowa	0-0	1007
10	Virginia Tech	0-0	973

   
AP 2010 Final Top 10 Poll 

1	Auburn (56)	14-0	1472
2	TCU (3)	13-0	1392
3	Oregon	12-1	1379
4	Stanford	12-1	1300
5	Ohio State	12-1	1220
6	Oklahoma	12-2	1108
7	Wisconsin	11-2	1055
8	LSU	11-2	1051
9	Boise State	12-1	1031
10	Alabama	10-3	961

Now let's look at where three teams that finished in the Top 5 that did not start in the Top 10, started in the preseason rankings. Auburn won the national championship and was ranked 22nd in the AP Preseason, Oregon finished third and started 11th, and Stanford finished 4th while starting unranked.
-----------------------------------------------------------------------------------------------------
Remarkably, out of the teams that started in the Top 10 in the 2009 season, seven finished in the Top 10. And of the teams to finish in the Top 5, four of them started in the Top 10.


AP 2009 Preseason Top 10 Poll

1	Florida (58)	0-0	1498
2	Texas (2)	0-0	1424
3	Oklahoma	0-0	1370
4	USC	0-0	1313
5	Alabama	0-0	1156
6	Ohio State	0-0	1113
7	Virginia Tech	0-0	1054
8	Ole Miss	0-0	1047
9	Oklahoma State	0-0	989
9	Penn State	0-0	989

AP 2009 Final Top 10 Poll

1	Alabama (60)	14-0	1500
2	Texas	13-1	1399
3	Florida	13-1	1370
4	Boise State	14-0	1366
5	Ohio State	11-2	1224
6	TCU	12-1	1163
7	Iowa	11-2	1126
8	Cincinnati	12-1	1060
9	Penn State	11-2	1016
10	Virginia Tech	10-3	953

---------------------------------------------------------------------------------------------------
Of the teams to begin in the Top 10 in 2008, only four finished in the Top 10. Three of the teams in the final Top 5 began the season in the Top 10. For the most part it seems as if the AP Poll does a good job of pegging a few elite teams, but misses on surprise teams and other teams that are simply overrated. Of the past three AP Polls, two teams from the Top 4 have played in the BCS Championship Game, meaning that in the past three years you had a greater chance of looking outside of the Top 4 to find a national title game participant than in looking in the Top 4. However, look at the Top 5, and in the past three years, 4 of the six participants have started in the Top 5. 


AP 2008 Preseason Top 10 Poll

1	Georgia (22)	0-0	1528
2	Ohio State (21)	0-0	1506
3	USC (12)	0-0	1490
4	Oklahoma (4)	0-0	1444
5	Florida (6)	0-0	1415
6	Missouri	0-0	1266
7	LSU	0-0	1135
8	West Virginia	0-0	1116
9	Clemson	0-0	1105
10	Auburn	0-0	968

AP 2008 Final Top 10 Poll

1	Florida (48)	13-1	1606
2	Utah (16)	13-0	1519
3	USC (1)	12-1	1481
4	Texas	12-1	1478
5	Oklahoma	12-2	1391
6	Alabama	12-2	1264
7	TCU	11-2	1193
8	Penn State	11-2	1153
9	Ohio State	10-3	1013
10	Oregon	10-3	997

Where Do National Championship Coaches Come From?

Looking at data from 1980, we find that from 1980 through 1986 all the coaches that won national titles had no previous head coaching experience. Some of those names are the biggest names in college football history, including Joe Paterno, Barry Switzer, LaVell Edwards, and Vince Dooley. In 1987, Jimmy Johnson started a trend reversal when he became the first coach with prior head coaching experience to win a national title since 1979 when Bear Bryant won his sixth and final championship at Alabama.

Larry Coker did not have previous head coaching experience.

Since 1987, 21 different coaches have won national titles, and of those coaches, 15 have had previous head coaching experience. From 1997 to 2001, five of the six coaches who won national titles did not have previous head coaching experience prior to being appointed head coach of their team. These coaches are Tom Osborne, Lloyd Carr, Phillip Fulmer, Bob Stoops, and Larry Coker. The lone exception was Bobby Bowden in 1999. All coaches to win a national title since 2002 have had previous head coaching experience, and only five coaches without previous head coaching experience have appeared in the BCS National Championship Game, including Bob Stoops, Phillip Fulmer, Larry Coker, Frank Solich, and Chip Kelly. Larry Coker was the last coach to win a national title without having previous head coaching experience in 2001. 


Though previous head coaching experience is preferable, especially in a day and age where coaches are highly unlikely to stay at a school for more than ten years, it is possible to find a highly qualified assistant that will create success at a program such as Stoops, Fulmer, Carr, and more recently, Oregon's Kelly.


A total of 27 coaches have won a national title since 1980. Of those 27, we found the coaching backgrounds of 25 of them.  Of those 25 coaches, 15 had offensive backgrounds and 10 had defensive backgrounds. 


1980 - Vince Dooley, no prior head coaching experience
1981 - Danny Ford, Clemson, no prior head coaching experience, offensive assistant
1982, 1986 - Joe Paterno, Penn State, no prior head coaching experience, offensive assistant
1983 - Howard Schnellenberger, Miami, Fla., no prior head coaching experience, offensive coordinator
1984 - LaVell Edwards, BYU, no prior head coaching experience, offensive coordinator
1985 - Barry Switzer, Oklahoma, no prior head coaching experience, offensive coordinator
1987 - Jimmy Johnson, Miami, Fla., prior head coaching experience, defensive coordinator
1988 - Lou Holtz, Notre Dame, prior head coaching experience
1989, 1991 - Dennis Erickson, Miami, Fla., prior head coaching experience, offensive coordinator
1990 - Bill McCartney, Colorado, no prior college head coaching experience, defensive coordinator
1990 - Bobby Ross, Georgia Tech, prior head coaching experience, defensive coordinator
1991 - Don James, Washington, prior head coaching experience, defensive coordinator
1992 - Gene Stallings, Alabama, prior head coaching experience, defensive coordinator
1993, 1999 - Bobby Bowden, Florida St., prior head coaching experience, offensive coordinator
1994, 1995, 1997 - Tom Osborne, Nebraska, no prior head coaching experience, offensive coordinator
1996 - Steve Spurrier, Florida, prior head coaching experience, offensive coordinator
1997 - Lloyd Carr, Michigan, no prior head coaching experience, defensive coordinator
1998 - Phillip Fulmer, Tennessee, no prior head coaching experience, offensive coordinator
2000 - Bob Stoops, Oklahoma, no prior head coaching experience, defensive coordinator
2001 - Larry Coker, Miami, Fla., no prior head coaching experience, offensive coordinator
2002 - Jim Tressel, Ohio St., prior head coaching experience, offensive coordinator
2003, 2009 - Nick Saban, LSU, prior head coaching experience, defensive coordinator
2003, 2004 - Pete Carroll, USC, prior head coaching experience, defensive coordinator
2005 - Mack Brown, Texas, prior head coaching experience, offensive coordinator
2006, 2008 - Urban Meyer, Florida, prior head coaching experience, offensive coordinator
2007 - Les Miles, LSU, prior head coaching experience, offensive coordinator
2010 - Gene Chizik, Auburn, prior head coaching experience, defensive coordinator

National Championship Winning Coaches

Bob Stoops has been a BCS mainstay.

When did coaches win their national championships at their school? I was interested in looking at the years in which coaches won a national championship to see if any trends would emerge. Since 1997, there has been an interesting trend emerge between coaches and the years they had coached at the school when winning a national title. I looked at data as far back as 1980 and included split national titles (1990, 1991, 1997, 2003). Below is a list of the national title winning coaches, their school, their championship and how many years they had been at the school, and how many years they coached at that school. My findings are below the list.

1980 - Vince Dooley, Georgia, 1st national title in 17th season at school (coached 25 seasons)
1981 - Danny Ford, Clemson, 1st national title in 4th season at school (coached 12 seasons)
1982 - Joe Paterno, Penn State, 1st national title in 17th season at school (entering 46th season)
1983 - Howard Schnellenberger, Miami, Fla., 1st national title in 5th season at school (coached 5 seasons)
1984 - LaVell Edwards, BYU, 1st national title in 13th season at school (coached 27 seasons)
1985 - Barry Switzer, Oklahoma, 3rd national title in 13th season at school (coached 16 seasons)
1986 - Joe Paterno, Penn State, 2nd national title in 21st season (entering 46th season)
1987 - Jimmy Johnson, Miami, Fla., 1st national title in 4th season (coached 5 seasons)
1988 - Lou Holtz, Notre Dame, 1st national title in 3rd season (coached 11 seasons)
1989 - Dennis Erickson, Miami, Fla., 1st national title in 1st season (coached 6 seasons)
1990 - Bill McCartney, Colorado, 1st national title in 9th season (coached 13 seasons)
1990 - Bobby Ross, Georgia Tech, 1st national title in 4th season (coached 5 seasons)
1991 - Dennis Erickson, Miami, Fla., 2nd national title in 3rd season (coached 6 seasons)
1991 - Don James, Washington, 1st national title 17th season (coached 18 seasons)
1992 - Gene Stallings, Alabama, 1st national title in 3rd season (coached 7 seasons)
1993 - Bobby Bowden, Florida St., 1st national title in 18th season (coached 34 seasons)
1994 - Tom Osborne, Nebraska, 1st national title in 22nd season (coached 25 seasons)
1995 - Tom Osborne, Nebraska, 2nd national title in 23rd season (coached 25 seasons)
1996 - Steve Spurrier, Florida, 1st national title in 7th season (coached 12 seasons)
1997 - Tom Osborne, Nebraska, 3rd national title in 25th season (coached 25 seasons)
1997 - Lloyd Carr, Michigan, 1st national title in 3rd season (coached 13 seasons)
1998 - Phillip Fulmer, Tennessee, 1st national title in 7th season (coached 17 seasons)
1999 - Bobby Bowden, Florida St., 2nd national title in 24th season (coached 34 seasons)
2000 - Bob Stoops, Oklahoma, 1st national title in 2nd season (entering 13th season)
2001 - Larry Coker, Miami, Fla., 1st national title in 1st season (coached 6 seasons)
2002 - Jim Tressel, Ohio St., 1st national title in 2nd season (coached 10 seasons)
2003 - Nick Saban, LSU, 1st national title in 4th season (coached 5 seasons)
2003 - Pete Carroll, USC, 1st national title in 3rd season (coached 9 seasons)
2004 - Peter Carroll, USC, 2nd national title in 4th season (coached 9 seasons)
2005 - Mack Brown, Texas, 1st national title in 8th season (entering 14th season)
2006 - Urban Meyer, Florida, 1st national title in 2nd season (coached 6 seasons)
2007 - Les Miles, LSU, 1st national title in 3rd season (entering 7th season)
2008 - Urban Meyer, Florida, 2nd national title in 4th season (coached 6 seasons)
2009 - Nick Saban, Alabama, 1st national title in 3rd season at school (entering 5th season at school)
2010 - Gene Chizik, Auburn, 1st national title in 2nd season (entering 3rd season)

From 1980 to 1989, there were 10 national titles awarded without a coach with repeat national titles. Five of the coaches won in their 5th season or less, and five of the coaches won a national title after more than 13 years at their school. Only Joe Paterno won a national title after more than 20 years on the job, winning his second national title in 1987 in his 21st season.

From 1990 to 1999, there were 13 national titles awarded, with one coach winning two titles (Bobby Bowden) and another winning three titles (Tom Osborne). Six national titles were won by coaches with more than 15 years at a school, with Osborne and Bowden accounting for five of those championships, and Don James accounting for the other. Four coaches won a championship in less than five years, and three coaches won a championship between their sixth and tenth years.

From 2000 to 2010 (11 seasons), there were 12 national titles awarded. Three coaches won two national titles (Pete Carroll, Urban Meyer, Nick Saban). Of the 12 national titles, not a single coach was at their school more than 10 years when they won. In fact, of the 12 national titles, 11 of them were won by coaches in their 4th season at a school or less. The lone exception was Texas' Mack Brown who won in his eighth season at the school.

Since 1997 - a span of 16 national titles - only two titles were won by coaches with more than 10 years at their school, and both coaches have now retired (Osborne and Bowden). Not a single coach has won between their 10th and 20th season coaching since 1993 when Bowden won his first national title in his 18th season at Florida St. Since 1986, only two coaches have won a national title between their 10th and 20th seasons at a school; the aforementioned Bowden and Washington's James who won in his 17th season.

I think this speaks for how competitive college football has become and how quickly a coach, given the resources, can comprise a national title team. I think it also speaks for the constant grind of the job and how difficult it is now to stay in coaching with one school like the Switzers, Dooleys, Bowdens, Osbornes, Fulmers, Paternos, Edwards' and James' of the coaching world. Of the teams in Phil Steele's preseason top 40, only seven coaches have been at their schools for 10 or more years. Perhaps newer coaches have an advantage over more experienced coaches in that they are innovating the way the game is being played.

Bob Stoops is now entering some rarified company as he begins his 13th season at Oklahoma, as is Texas' Brown as he enters his 14th season. For whatever reason, however, the odds are stacked against them that they will win a title between now and their 20th season, which makes sense since the field is stacked against them.

I would give Stoops the best chance of becoming the first coach with more than 10 years of experience to win a title since Bobby Bowden did it in 1999 in his 24th season. Stoops has become one of college football's innovators which has allowed him to appear in four BCS Championship Games in 12 seasons.

Job Security of College Football Coaches; Introduction of "Win Gain"

In doing some research of college football coaches, I came across an interesting stat suggesting the "fireability" of a coach through four years. From 2000-2010, there were 113 coaches hired at BCS conference schools and at five additional schools; Notre Dame, BYU, Utah, TCU, and Hawaii. I looked at the first four years because that is typically the amount of time that a school will give a coach to prove their worth. In fact, to find how many times a university will give its coach four years, we reduced the 113 hires by the number of coaches who left for another job before their fourth year at a school, willingly resigned prior to their fourth year, and active coaches that have not or are not yet preparing for their fourth season at a school. The research indicated that 85 coaches did not leave for another job, are at least entering their fourth year at a school, or were fired before their fourth year. Of those 85 coaches, a total of 72 -- or 84.7 percent -- of their schools allowed them to enter their fourth year.

Stanford's Jim Harbaugh compiled the highest "net win gain."

Now, I came across a stat I'll call "win gain." The "win gain" takes a look at the records of five seasons, the first record is the record the departing coach left an incoming coach. The next four records were a coach's first four seasons. Some coaches did not last four years, or even three years. So a coach with four seasons would have had five seasons of comparison, a coach with three seasons would have had four years of comparison, a coach with two years at a school would have three years of comparison, and a coach with one year at a school would have two years of comparison. The net win gain suggests that in the first four years an incoming coach is measured against the final record of his predecessor and by the consistency and success of his own tenure. A coach can build a lot of goodwill in his first four years at a school, and based on the previous stat it shows that the majority of schools are willing to give coaches the four years to do that.

Let's take a look at former Minnesota coach Tim Brewster as an example. The departing coach, Glen Mason, left his final season at Minnesota with a 6-7 record. Brewster finished his first season with a 1-11 record, leaving him with a net "win gain" of -5. Ouch. Now, in Brewster's second season, his team finished with a 7-6 record, leaving him with a "win gain" from the previous year of +6. After two seasons, Brewster had a "net win gain" of +1. Brewster's third year he went 6-7, and then in his fourth year he was fired midseason after tumbling through a 1-6 record. Brewster coached into his fourth year, giving us five years of comparison that includes his predecessor's final season, which I will now list: 6-7, 1-11, 7-6, 6-7, 1-6. Brewster's net "win gain" in the four seasons he coached was -5.

Four important disclosures, 1) I did not include the records of interim coaches who only coached a few games of a season or one bowl game, but I did include the team's final record as the basis of comparison for the first season of an incoming coach, regardless of the fact that two coaches were in charge of a team during one season. (ex: Brewster went 1-6, then interim coach Jeff Horton went 2-3, so the comparison for incoming coach Jerry Kill will be a 3-9 season.) 2) Losses are irrelevant in this calculation. What we are looking for is an increase in wins, which is actually favorable to a coach since schools have now moved to a 12-game schedule, thus giving coaches more opportunities to win more games. 3) In my opinion, you can only really fairly use the net win gain after a coach's fourth season, though we did extend it for coaches that have coached through their third season for data-gathering purposes. 4) The data provided gives a snapshot of today entering the 2011 season, not of any trends from previous seasons.

What I found was startling. Though I'm certain that this calculation has many flaws, it indicated which coaches were likely to be fired after four seasons, and which ones were likely to keep their jobs. The overwhelming majority of coaches with a positive net win gain were retained, while a slight majority of coaches with a negative net win gain were fired. And coaches that had a neutral net win gain, or a 0, had a good chance of keeping their jobs. Now you would say that this makes sense, why would a school keep a losing coach? And it's obvious that a school would keep a winning coach. But that's not necessarily the case. For example, let's take a look at Vanderbilt coach Bobby Johnson who coached for eight seasons at Vanderbilt.

Since Johnson coached for eight seasons, we have five seasons of comparison for the net win gain, the first being the record left to him by his predecessor. Comparing his predecessor's last season with Johnson's first four years we have: 2-9, 2-10, 2-10, 2-9, 5-6. His net win gain after four seasons was +3. He kept his job and went on to coach four more years before retiring just before the 2010 season. You say, that's Vanderbilt, they don't really care about football there so Johnson was given more time to turn things around. Besides, he probably saved his job because he had a much better fourth year, which is absolutely true. The importance of a fourth year cannot be underestimated.

Rich Rodriguez could have benefitted from a 4th year

But let's go look at a school that really does care about its program, Michigan. Rich Rodriguez coached at Michigan for three seasons before being fired, all while showing progress in the win column in his three years there. Let's calculate his net win gain through three seasons: 9-4, 3-9, 5-7, 7-6, net win gain of -2. After three seasons Rodriguez was fired, and I would argue that had he won nine games in his third season, despite all the controversy surrounding him, he would have kept his job, or that he would have needed to win at least nine games in his fourth season if he had any shot of keeping his job. But Rodriguez did not win nine games in his third season, and he wasn't given the chance to enter a fourth season, but we do know where his net win gain stood after three seasons.

There were 82 coaches hired between 2000 and 2008 that have coached at least three seasons at their schools. Of those coaches, we could exclude Les Miles (OK State), Jim Harbaugh (Stanford), Mark Dantonio (Cincinnati), Brian Kelly (Cincinnati), and Bobby Petrino (Louisville) because they did not keep their jobs by accepting other jobs after either their third or fourth seasons at their school. But it is safe to assume that each one would have been back at their post if they had the desire to do so. Each one was extremely successful at their school and had a positive net win gain at the time they left their school. So we will stick with a sample of 82 coaches who coached at least three seasons at their school regardless of whether or not they left for another job.

Did Ty Willingham deserve a 4th year at Notre Dame?

Of the 46 coaches with a positive net win gain after either their third or fourth seasons, all have kept their jobs with the LONE exception of Ty Willingham during his tenure at Notre Dame. Keep in mind, this does not mean the coach was not fired after their fifth or subsequent seasons, but that the university allowed them to go beyond a fourth year. Coaches on this list include (Sylvester Croom (5 total seasons, 17-30 in first four seasons), Rich Brooks (7 total seasons, 17-30 first four seasons), Bobby Johnson (8 total seasons, 11-35 first four seasons), and Charlie Weis (5 total seasons, 29-21 first four seasons). Former North Carolina coach Butch Davis, though fired just before the beginning of his fifth season, does not count against this list because he was not fired for on-the-field performance. After three seasons at Notre Dame, Willingham's net win gain was +1, compared to Weis' +2 after four seasons.

Of the 13 coaches with a negative net win gain after four seasons of coaching, six kept their jobs [Les Miles (LSU), Larry Coker (Miami), Bill Doba (Washington St.), Dennis Erickson (Arizona St.), Dan Hawkins (Colorado), and Gary Patterson (TCU)]. It should be noted that Miles and Coker had won national titles within their tenures, and Erickson had won two national titles while at Miami. Doba and Hawkins were fired after their fifth seasons, Coker was fired after six seasons, Miles has been a controversial coach at LSU, Erickson should be considered on the hot seat entering this season, and Patterson has become one of the best coaches in the country.

Of the 12 coaches with a negative net win gain after three seasons of coaching, five kept their jobs [Dabo Swinney (Clemson), Paul Johnson (Georgia Tech), Paul Wulff (Washington St.), Rick Neuheisel (UCLA), and Greg McMackin (Hawaii)]. Those that lost their jobs are Ed Orgeron (Ole Miss.), Ron Zook (Florida), Gerry Di Nardo (Indiana), Bobby Williams (Michigan St.), Rich Rodriguez (Michigan), Buddy Teevens (Stanford), and Steve Kragthorpe (Louisville).


Of the 10 coaches with a neutral net win gain after four seasons, seven kept their jobs [Jim Tressel, Bret Bielema, Ralph Friedgen, John Bunting, Chan Gailey, Karl Dorrell, and Rich Rodriguez (West Virginia)] The three that were fired were Randy Shannon, Bill Lynch, and John L. Smith. 

There was one coach, Ron Prince (Kansas St.), with a neutral net win gain after three seasons, and he was fired after the 2008 season.

Which coaches had the highest net win gain after four seasons? Jim Harbaugh (Stanford) +11, Nick Saban (LSU) +10, Pete Carroll (USC) +8, Jeff Tedford (Cal) +7, Mark Dantonio (Michigan St.) +7, Sylvester Croom (Miss. St.) +6, Urban Meyer (Florida) +6, and Tom O'Brien (NC State) +6.

Which coaches had the lowest net win gain after three or four seasons? Steve Kragthorpe (Louisville) -8, Bobby Williams (Michigan St.) -7, Tim Brewster (Minnesota) -5, Buddy Teevens (Stanford) -5, Bill Doba (Washington St.) -4, Paul Wulff (Washington St.) -4, Bill Callahan (Nebraska) -4, Ted Roof (Duke) -4, Dan Hawkins (Colorado) -4, and very surprisingly Gary Patterson (TCU) -5. Patterson's negative win gain is the result of having an inaugural 6-6 season after a 10-2 season by Dennis Franchione, and following up an 11-2 season in his third year with a disastrous 5-6 season his fourth year.

Disturbingly, where 46 coaches had a positive net win gain after their third or fourth seasons, only Willingham was not retained. Of the 11 coaches with a neutral net win gain after three or four seasons, only four were not retained, including Miami's Shannon and Kansas State's Prince. All the coaches mentioned above are African-American. It's not to say that their firings didn't have some merit, but their firings are also contradictory to what the numbers say.

I totally made up this "win gain" stat and it may very well be that it has no significant bearing on anything, but it is interesting to see where the exceptions lie and where some trends emerge. I know for sure that I am not the first to think of the concept of a "win gain" or to apply such a concept over a prescribed amount of time (in this case 3 and 4 years). Perhaps athletic departments take this idea into account, whether subconsciously or not, when deciding the fate of their coach, or not. It'll be interesting to track this little stat as more coaches finish up their third and fourth seasons.

College Football Coaches in Their 2nd Year

For the first time ever I bought a preseason magazine for college football, Phil Steele's annual preview. I was impressed by all the work he put into it and all the numbers he crunches to come up with what is ultimately an educated guess.

Jimbo Fisher (left) has rightfully high expectations at FSU.

In that same vein, I decided to crunch a few numbers of my own relating to coaches at BCS conference schools and five additional programs (Boise St., TCU, Notre Dame, Hawaii, and BYU). I took a look at coaches hired from 2000-2010, giving me 11 years of data. Some coaches are still active, some just finished their first year in 2010, and others have been hired more than once. 

The first stat I wanted to look at was how much coaches improved their school's win total from one year to another. I only took a look at the first four years of a coach's tenure, realizing that some coaches did not make it to even their second year at a school, let alone their third and fourth years, and a small number of coaches are at their same school after 10 years or more. Schools typically give a coach four years to work on their program before deciding on whether to keep him or not. 

Out of 113 coaches hired at the schools I studied, 49.5 percent were able to increase their school's win total in their first year, 13 percent stayed the same, and 37.5 percent went down. Remarkably, of the 100 coaches that have made it to their second year, 59 percent increased their win total from the previous year, 13 percent stayed the same, and 28 percent went down in win totals. But things take a downward turn in the third year. 

Of the 82 coaches in their third year at the same school, only 41 percent increased their win totals from the previous year, 12 percent stayed the same, and a whopping 47 percent went down in win totals. I figured that this would be the case if a coach had a strong second season, then it would be difficult to match those win totals in the third year. OK. I can buy that argument. So I went ahead and took a look at the coach's fourth year, a critical year for all coaches in terms of contract extensions and retainability. Of the 60 coaches in their fourth year, 45 percent increased their win totals from the year before, 5 percent stayed the same, and an eye-opening 50 percent went down in win totals.

This speaks to the importance of a coach's second season. This information doesn't speak to any one school looking at insight for what to expect from their coaches, but it does appear that coaches are able to build upon their first year 59 percent of the time, and at no other point in their first four years will they be more likely to increase their school's win total from the previous year than in that second year. This means good things, especially if your coach's first season was a good one, because the likelihood your school will not at least match the previous year's win total is only 28 percent. 

I think the numbers also suggest that being a college head football coach is very difficult, and to be able to sustain a winning program is even harder. In some ways, the longer a coach is at a school, the harder it may become to continue building off of any initial success. It could also suggest that a coach simply isn't very good or isn't a good fit at a school, and it takes about four years to finally figure that out. 

This is by no means a guarantee of anything in any way, but it speaks well of a second-year coach in charge of a talented team. Recent second-year coaches that have won a national title or appeared in the title game since the 2000 season include Bob Stoops, Urban Meyer, Chip Kelly, Jim Tressel, and Gene Chizik. If this trend holds up, Notre Dame's Brian Kelly and Florida State's Jimbo Fisher should very well  be considered legitimate national title contenders.