|
|
|
|
|
|
|
Jens 'n' Frens
Idle thoughts of a relatively libertarian Republican in Cambridge, MA, and whomever he invites. Mostly political.
"A strong conviction that something must be done is the parent of many bad measures."
-- Daniel Webster
|
|
|
|
|
|
Monday, January 06, 2003 :::
In its five-year history, the BCS (college football) championship has gone to a different conference each year. (If you suppose each conference to have an equal chance of winning each year, uncorrelated one year to the next, the odds of that are 5/54, about 9.3%.) Some thoughts on championships in general, and college football in particular:
- The Fiesta Bowl was a great game, but I found myself disappointed that a championship game wasn't more decisive. Even a decisive — say, double-digit — win by a worse team against a better team is not that uncommon in football, but at least it feels like a data point. This game told us that the teams are probably pretty evenly matched. If the purpose of the championship game is to give us a great game between two great teams, this did a good job of that. Determining the better team it didn't.
- The whole idea of a championship — and I myself make this assumption elsewhere in this post — assumes a kind of transitivity that is certainly an approximation to reality. A can beat B can beat C can beat A on a repeatable basis, just based on which teams run well, pass well, etc. To the extent that this is false, we kind of have to define what we even mean if we're striving to find out which team is the best.
- Of course the better team doesn't always win, even in a decisive game — in the NFL, an undefeated season is once heard of — and that the world of college football seems a bit unduly interested in undefeated teams. The estimable Jeff Sagarin put USC #1 in his final rankings, both those using his traditional system and those prepared for the BCS that ignore margin of victory; this is certainly an interesting claim of a team that lost two games, but is by no means indefensible.
- To emphasize that point: two of the teams that won NFL playoff games this weekend lost to the Chicago Bears this season. (One of them beat a team that beat the Bears twice.)
- I've always been skeptical of the idea of a (large) play-off tournament because of the substantial chance for upsets and the reliance on a small number of data points (as opposed to a whole season), but insofar as the playoff system encourages teams to produce tougher schedules without fear of a single loss eliminating them — in college basketball, top teams routinely acquire several losses early in the season without excessive concern — it could yield a system that, taken as a whole, does result in a higher probability that the actual best team will win the championship.
- I do like my —s.
- I'm not a fan of overtime in general; if a game ends in a tie it really isn't decisive whether it's left that way or broken by means of some perversion of the sport in question. (Basketball overtime doesn't pervert the sport so badly, but football, both college and professional, do. Incidentally, I went to bed Friday night after regulation.) There are situation in which a winner simply has to be chosen — single-elimination playoffs, for obvious example. When one doesn't, I don't think the win-loss record should be drained of some of the little information it has in it.
::: posted by dWj at 2:26 PM
|
|
|
|
