Stewart Mandel’s latest Mailbag contains a chart that I find very interesting. The chart lists three measures of strength of schedule for every team that has played in a BCS Championship Game. The results are fairly telling. Auburn has played the third-toughest schedule, while Oregon’s is dead last. Now, it’s important to mention that simply looking at the records of a team’s opponents isn’t a great way to measure strength of schedule. Specifically, it penalizes a team from the Pac Ten because the Pac Ten has a nine-game conference schedule, which will tend to push the records of Oregon’s opponents towards .500. SEC teams play eight conference games and tend to replace that ninth conference game with a revenue-friendly tomato can, which inflated the overall record of Auburn’s opponents. Massey and Sagarin show Auburn and Oregon as having played equivalent schedules; SRS shows Auburn has having played a tougher schedule, but the margin isn’t huge. I’d bet that Jerry Palm himself, the person that Mandel refers to has “our hero,” would be unimpressed by the methodology that Mandel uses.
That said, look at what has happened in title games where one opponent was more than ten spots removed from the other on the chart:
1998 Tennessee (20) over Florida State (1)
2006 Florida (4) over Ohio State (24)
1999 Florida State (5) over Virginia Tech (25)
2009 Alabama (7) over Texas (19)
Three of the four teams that played significantly tougher schedules won the title game and the fourth is a bit of a misnomer because Florida State was playing without its starting quarterback. The Noles wouldn’t have been in the title game in the first place if Marcus Outzen would have been their quarterback for the whole season. (Counterpoint: if we drop the ‘98 game, then we should also drop the ‘09 game because of Colt McCoy’s injury and then we’re left with a sample size of two.)
The more I think about it, the more I’m deciding that Mandel has a good point about strength of schedule, but he’s using the wrong stats to achieve it.