One of the interesting issues that has emerged after the Willingham hire has been how much of a team's performance to attribute to coaching and how much to the raw talent of the players. Of course, most everything is attributable to the coach ultimately, as he is responsible for recruiting the players. But the years immediately preceding and succeeding a coaching change are a good opportunity to evaluate how much can be chalked up to doing the most with the talent available -- conditioning, game planning, scheme, play calling, putting players in the right positions and so on -- and how much is attributable to the raw talent of the team. In a coach's first year (and to a lesser degree his second) he can't do much about the talent level of the team on the field. Unless he recruits some impact freshmen (rare, to say the least) there's essentially nothing he can do about it his first year. That's not to say that the talent level doesn't change, but it has more to do with the players leaving than those coming.

I decided to see if any statistical analysis could help shed light on this question. As usual, this comes with my standard warning that if you don't like the numbers stuff, avert your eyes.

In some of my recent musings, I've gotten interested in a team's total points scored for the season and total points allowed. In a certain sense, this is a truer measure of a team's dominance than raw record.  Davie generated very weak results by this measure, even in his relatively good years (1998 and 2000). So I went back to three years before Ara started and collected the total points scored for the year and total points allowed and calculated the differential, i.e., the total margin for the season.

Next to the year, I note if either a coaching change occurred that year or if N.D. won the national championship or came close (defined as no more than one loss). Here it is, year by year:

1961: - 7
1962: - 33
1963: - 51
1964: +210 Near National Championship (1 loss), Ara hired
1965: +197
1966: +324 National Championship
1967: +213
1968: +206
1969: +217
1970: +246 Near National Championship (1 loss)
1971: +139
1972: + 97
1973: +293 National Championship
1974: +171
1975: +100 Devine hired
1976: +136
1977: +287 National Championship
1978: + 96
1979: + 46
1980: +120
1981: + 72 Faust hired
1982: + 32
1983: +139
1984: + 60
1985: - 4
1986: + 80 Holtz hired
1987: +121
1988: +237 National Championship
1989: +238 Near National Championship (1 loss)
1990: +100
1991: +176
1992: +246 Near National Championship (1 loss, 1 tie)
1993: +212 Near National Championship (1 loss)
1994: + 62
1995: +145
1996: +226
1997: + 17 Davie hired
1998: + 80
1999: + 17
2000: + 86
2001: - 1

At a macro level, it's easy to see that the changes correspond -- and correspond immediately -- with our instincts about who was a good coach and who was not. Ara was hired, and the point differential improved by 261. Davie was hired, and it took a 209 point loss (Davie: the anti-Ara). Holtz was hired and it improved by 84. Faust was hired, and it took a 48-point hit. Devine -- a good coach, but not Ara -- was hired and it lost 71 points.

The absolute levels are quite interesting as well. It takes a differential in the low 200's to compete for the national championship, close to 250 to have a good chance to win it. Interestingly, these differentials are not limited to N.D. Although other schools (esp. with weaker schedules and higher octane offenses) have generated some much larger differentials, the low to mid-200's seems to be a level necessary to get you in the derby. Penn State's 1982 team was +199, its 1986 team was +207, but its 1994 team was +312 and lost out  -- despite being undefeated -- to Nebraska.. Nebraska's 1994 team was +297, its 1995 team (the one that threw Fla. around like rag dolls in the Fiesta Bowl) was an astonishing +472 and its 1997 team that shared it with Michigan was +393. (Solich, by the way, has not come close to Osborne's levels). Michigan that year (1997), though, was a more modest +208. The Florida St. team that beat out ND in 1993 was a huge +407 (even with a negative 7 to N.D.) and its 1999 team was +255. Tennessee in 1998 was +242 and Oklahoma last year was +287. You get the idea: it takes at least +200 -- assuming you're playing a major conference schedule -- to get you sniffing around the edges of the championship race, 250 puts a team like N.D, Oklahoma, Penn St., Michigan, etc. seriously in the hunt.  To get in the hunt with a much lower differential requires a lot of luck.  Luck can go the other way too, as Holtz's biggest differentials didn't win him the championship; ditto for Paterno. Some of it is even more random than this. Devine had the good fortune to lose to Mississippi early in the year in 1977, Holtz the bad luck to lose late in 1993 and to Miami in 1989.

Now, how much to chalk up to talent and how much to just raw coaching? In any year in which there is NOT a coaching change, the average differential (the absolute value of the delta for propeller heads) is 73.6.  In a year in which there's a coaching change, the average differential (again, either way -- the absolute value) is 134.6.  The differential in non-coaching change years has to be put down to different levels of talent, changes in the schedule, injuries and the like.  But the considerably higher differential in coaching change years has to be put down to the differences in coaching ability, conditioning, preparation, game plan, play calling, clock management and the like.

It's actually a little easier to see if you look two years either side of a coaching change. In two years, unless he's able to play a lot of sophomores, a coach still hasn't had much of a chance to make a positive impact on the overall talent level. In the two years before Ara, the margin averaged - 42 and in the two years after +204. In Ara's last 2 years it averaged +234 and in Devine's first two it was +118. In Devine's last two years it was +83 and in Faust's first two it was +53. In Faust's last two years it was +28 and in Holtz's first 2 it was +101. In Holtz's last two years it was +181 and in Davie's first two years it was +47. In Davie's last two years it was +42, and in Willingham's first two years it . . . Well, that chapter has yet to be written.

In the end, I'd say it's about a 50-50 mixture. Holtz and Ara, mostly on coaching skill alone, were able to get the team roughly halfway (in Ara's case more than halfway) to where it needed to be to win the national championship. By the time they played mostly with their own players in years 3 and 4, they dominated. Devine was more of a maintainer: he kept the program at a reasonably high level post-Ara, had a knack for winning close games, and put together one terrific season. Faust was incompetent, but maybe not quite as badly as Davie (this, again, is a very close question), but managed to drive the program into the ground. Davie had an even more sudden negative impact than Faust, but it was tragically disguised a bit by a couple of lucky 9-win seasons. If Willingham gets us around +100 (and he got Stanford there this year in the regular season), we are definitely headed in the right direction.