Fuzzy Math: How College Football's Champs Were Chosen
The BCS pairings have been announced for this year’s football season, so here’s a question for you college football nuts: what NCAA school claims the most football national championships?
Who did you guess? Alabama? The Crimson Tide have only won a paltry 13 titles. Michigan? Notre Dame? USC? They’re all tied with a pedestrian 11 championships. No, there’s only one major championship owner: Princeton. Yes, that Princeton. The Tigers claim 28 championships, which gives them a slight edge over Yale’s 26 titles. Bama’s 13 titles makes it a rather distant third. The lesson: if you want real football tradition, head to New Jersey. If that doesn’t work, you can always settle for Tuscaloosa or South Bend.
Of course, as you might expect, these numbers are a bit misleading. Neither Princeton nor Yale has hoisted a title since Princeton’s 1935 campaign, and back in those days “national title” didn’t mean quite the same thing that it does now.
Over the years, upwards of thirty groups have used polling, statistical formulas, historical research, and other methods to pick “national champions.” Some of these systems have been more rigorous than others, and it’s at a school’s discretion to decide whether or not it wants to claim a national title from a particular group.
As one might guess, the lack of a unified national title has led to some odd situations. For many seasons, the various systems couldn’t come to anything close to a consensus, and some years their conclusions were almost comically different. Take 1921, for example. Depending on which poll or system you consult, the national champ was either Cal, Cornell, Iowa, Lafayette, or Washington & Jefferson. Sort of puts our current “lack of a true champion” in perspective.
The polls that we rely on to crown our champions today have only been around since the AP poll debuted in its current incarnation in 1936. The Coaches’ poll followed in 1950. Since the advent of the polls, the list of teams that have won the most championships looks pretty much like you’d expect. Alabama and Notre Dame are tied for the top spot with eight outright or shared poll titles apiece, and Oklahoma and USC are right on their heels with seven each.
They Played to Win the Games (Not National Titles)
Modern playoff proponents consistently decry the use of the poll system for choosing the national champion, but things could be so much worse. For many years, there wasn’t a national champion at all. Rutgers and Princeton met in the first intercollegiate “football” game in 1869, but it didn’t look much like the gridiron action we’re used to seeing. In that game, 25-man sides tried to kick a ball into their opponents’ goal, all while refraining from carrying or throwing the ball. (Rutgers won 6-4.)
Although that clash would look odd to modern fans, it’s considered the first real college football game. The game gradually evolved into what we know as football, but schools weren’t really interested in claiming a national title. Teams just played each other as an end unto itself. In fact, it doesn’t seem like anyone put much thought into crowning a champion until the 20th century began.
In 1901 journalist Caspar Whitney began conducting a national poll to compare teams from different parts of the country. Whitney was something of an expert in college football; he had collaborated with Yale coach Walter Camp to select the first All American team in 1889. By 1905, Whitney’s poll was selecting a national champ at the end of the season. Sure, the poll only lasted through the end of the 1907 season, but it was a start.
The Numbers Game
The much-reviled computers of the BCS had forerunners of their own. In 1926 University of Illinois economics professor Frank G. Dickinson debuted the first mathematical formula for ranking the nation’s college football teams and declaring a champ. The Dickinson System was pretty simple. It divided teams into two groups: weak and strong, and it awarded schools 30 points for beating a strong team and 20 for beating a weak team. Defeats were worth half as much as a victory, so falling to a strong team still earned you 15 points. Ties were worth the average of a win and loss. Once Dickinson calculated all the scores, he then found each team’s average score per game. Bingo! Stanford won the 1926 title under the Dickinson System.
Dickinson’s system kept awarding national championships through the 1940 season. Notre Dame coach Knute Rockne grew curious about Dickinson’s formula and asked the professor to calculate scores for the 1924 and 1925 seasons that came before the formula’s debut. Further evidence that Rockne was a coaching genius: the legendary “Four Horsemen” Irish team came out on top in Dickinson’s 1924 rankings.
Oddly, even thought Dickinson’s system and the accompanying Rissman National Trophy were considered big deals in their day, USC managed to forget that it won one over the years. In 2004 the school declared that it had been a bit remiss in its record-keeping and for decades had neglected to claim the 1939 national title the Dickinson System had awarded the Trojans. USC had a ceremony at halftime of a game during its 2004 season and amended its records to show an additional national title.
You Were the Champions
If Dickinson’s method seems like a back-of-the-envelope calculation, the system Parke H. Davis debuted in 1933 looks charmingly quaint. Davis had played for Princeton in 1889 and later coached at Wisconsin, Amherst and Lafayette. Following his coaching stints, Davis became one of the leading early football historians. In 1933 he realized the value of awarding a national champion every year and wanted his say in the matter, so he sat down to retroactively award national titles throughout college football’s history.
Davis didn’t really have a methodology. According to a 1967 Sports Illustrated story on Davis’ project, “He used no special formula. He simply looked at the schedules and the results and chose his teams.” Davis didn’t mess around about going all the way back through football history, either. The 1869 “season” where Princeton and Rutgers played the aforementioned soccer-like game? They shared Davis’ national title that year. (Princeton had won a rematch played under a different set of rules. The schools’ faculties canceled a proposed rubber match because it was interfering with both teams’ studies. Both teams finished the year with 1-1 records.)
Davis’ project, which debuted in the 1933 edition of Spalding’s Football Guide, had another fun quirk: it put “national championship winning coach” on the resume of…Parke H. Davis. After doing his research, Davis declared Princeton and the Lafayette team he had coached to be co-champions of 1896. He wasn’t being self-aggrandizing, though. His squad had gone 11-0-1 while scoring 240 points and only giving up 10. Their lone tie was a 0-0 affair with co-champion Princeton.