Ah, yes, I live in Oklahoma. And most of my co-workers are very much into college football. I, for one, enjoy watching the game, but don’t have a particular attachment to any team. But since most people I know do so, I tend to root for OU (I suppose then, if I have a bias in this article, it’s in favor of OU). So, anyway, naturally I’ve heard some of the recent debate/fiasco over the Big 12 method of deciding the division championship in the case of a round-robin three-way tie. I actually have a few comments on that method as opposed to the SEC method (eliminate third in the standing, then take the head-to-head winner of the other two) as I’ve listened to the discussions.
Problems with Elimate-the-Last method
My understanding of the SEC method may be incorrect, as I’m not interested enough in this thought problem to actually research it, so I’m examing what I understand to be the process (as noted parenthetically above). It has some problems.
- The team that has the best (point wise) round-robin victory is very likely to lose the division championship, and the team with the least impressive point wise victory is likely to win. Because the team with the most significant loss (unless all games were pretty close) is probably going to be ranked third in the polls of those teams, and it’s likely that loss was to the top-ranked team. Guess who wins the division? The team not involved with the most lopsided game. It’s not guaranteed, but it’s likely.
- A team is punished for winning a lopsided victory. Because the loser is of that game is likely the one to be elimated (in essence, this is just a different angle of considering point #1).
- In a round-robin tie, there’s no way to logically use head-to-head victories as a metric. They’re all in the same boat. Eliminating a team using polls is not inherently different than choosing a winner based on polls (The poll voters essentially choose the winner in either case).
So, I think the SEC method is problematic, but what about the Big 12 method. Well, obviously, it’s not satisfactory to everyone. But at least it leaves the voters and computers trying to choose the best team of the three instead of the worst. The reality, though, is there’s just not enough games to come up with a non-controversial strategy. Some 120 teams with each team playing only 10% of the others. I imagine it’s simply not possible to effectively judge between them all. Even within a division, the teams play each other only once. There’s just enough games to ensure that every weakness will show up once, but not enough games to determine which weaknesses are general and which are one-time.
To me, the most sensible approach to a round-robin tie is point spread. Compare the margin of each team’s victory minus the margin of it’s defeat. After that, let the polls decide.
But, it must be acknowledged that determining clear champions in college football is a losing battle.