Monday, 17 November 2003

What I'd do to the BCS

As it’s Monday, it’s time for the weekly howls of outrage to erupt at the latest BCS standings. Unfortunately for fans of college football, however, the outrage is largely manufactured and misplaced. Why?

  • Controversy sells. Getting people to watch the 6 EST SportsCenter is pulling teeth; hence why ESPN has pulled Dan Patrick back into full-time duty in Bristol, and why the BCS standings are a prominent part of the Monday show—to the point that they receive nearly 48 hours of pre-hype from “College GameDay Final” on.
  • The controversy is manufactured by the entities that have the most to lose from an independent evaluation of college football: the media. 65 of America’s leading college football writers and broadcasters have a vested interest in their ratings being the sole indicators of quality in college football. The regional and other biases of both the writers and the coaches are notorious. Nothing like some diversionary controversy to deflect attention away from the gorilla sitting in the corner.

There are legitimate reasons to critique the BCS standings. The fundamental problem is that they’re an ad hoc amalgamation of polls, an arbitrary selection of computer rankings, and fudge factors, necessitated by the false legitimacy that the Associated Press and ESPN/USA Today polls have among college football fans. From an econometric standpoint, there are serious problems with the BCS.

A fundamental problem is that truly ordinal data is treated as metric in the formula. Your age, height, and weight are metric data: differences in age have real meaning. If I’m 27 and my cousin is 3, the difference in our ages—24 years—is a meaningful quantity. By contrast, poll rankings aren’t metric. LSU is #3 in the AP poll, and Ole Miss is #15. 15-3=12. Twelve doesn’t tell us much of anything about the difference between LSU and Ole Miss; it just tells us that there’s a difference. Missouri is #27. 27-15=12. Treating this difference as metric makes an invalid assertion: that the difference in quality between LSU and Ole Miss is the same as the difference between Ole Miss and Missouri.

This problem repeats itself throughout the BCS formula. Means of rankings in polls and computer rankings are taken. These means are added together. The strength of schedule component—which is a key component of many computer rankings—starts as metric data, then is converted to a ranking and arbitrarily scaled… then added to the means. Losses—which are metric—are then subtracted. Finally, an ad hoc adjustment is made for so-called “quality wins”—an adjustment one would hope that is incorporated in the polls and computer rankings anyway. Then the rankings are reported with these bizarre totals attached, apparently because totals look cool (I guess they got the idea from the AP and ESPN polls, who report the sorta-kinda metric Borda count in addition to the rankings).

Nonetheless, the fundamental idea of the BCS rankings is sound, even if there are too many compromises and too many ad hoc adjustments. So what would I do?

  • Include more computer rankings.
  • Use averaging methods appropriate for ordinal data. Or at least, recognize that taking the mean of a bunch of ordinal data doesn’t make it metric… so make it properly ordinal again.
  • Eliminate the silly restriction that computer rankings cannot incorporate margin-of-victory as a factor in their formulas. (I’ll explain why this restriction is silly in another post.)
  • Eliminate the ad hoc adjustments.

Next time (which I intended to be this time—sigh), I’ll talk about “computer rankings” in more detail. It turns out that they can be thought of as an application of the oft-maligned statistical technique known as factor analysis.


Mark your calendars… December 2nd is the day, at a secure, undisclosed combined conference room/classroom somewhere in Deupree Hall on the University of Mississippi campus. Of course, that’s not exactly the end of the tunnel, but pretty darn close.

Radical Interpretation of Matthew Yglesias

Up until today, if I had been asked to name the blogger that I most agree with (not necessarily my favorite blogger), it would have been Matthew Yglesias. This shouldn’t be surprising: we both have a background in analytic philosophy, both fans of David Lewis, we’re both consequentialists, we’re both liberals, and we’re both proponents of free trade.

But today he’s said something so outrageously false that, like Donald Davidson’s hypothetical man who says “There is a hippopotamus in the refrigerator” (from "On Saying That"), I have to wonder whether I’ve misinterpreted him.

Blogging about this list of the top ten albums of all time from Rolling Stone magazine, Matthew writes:

I would suggest that if you come to the conclusion that The Beatles are responsible for four of the top ten albums of all time, then your methodology is probably a bit off (they’re not, after all, the best band by whole orders of magnitude).

At first glance, he would seem to be saying here that not only are the Beatles not the best band of all time, they’re not even in the top ten.

But since this is self-evidently false, I must excercise Davidson’s "principle of charity". Like Davidson’s man who says “Look at that hansome yawl” while pointing at a ketch ("On the Very Idea of a Conceptual Scheme"), I must conclude that Matthew uses the term “Beatles” to refer to some group other than John, Paul, George, and Ringo. Perhaps he’s confusing them with the Monkees.

I must, however, agree with Matthew that Rolling Stone’s methodology must be a bit off. Not for the reason that he cites, but because Abbey Road is not one the four Beatles albums they put in the top ten.

Update: Brian Weatherson, resident philosopher at Crooked Timber, weighs in with his top ten list. His methodogy is a bit off too, since he also leaves out Abbey Road.

Just for fun, here's my top eleven list (I just couldn't winnow it down to ten), in no particular order, except Abbey Road is at the top:

Broker THIS!

Steven Taylor throws cold water on the idea that the Democrats will have a so-called “brokered convention”—i.e. that the plurality winner of the primary process won’t be the ultimate nominee. This isn’t the 1960s, and the Democratic base—particularly the Deanites—isn’t going to accept such meddling from party elites, and no amount of wishful thinking from either the media or anti-Dean forces in the party is going to affect that.

To get someone—anyone—other than Howard Dean as the nominee is going to require a lot of anti-Dean Democrats to swallow their pride and put the party ahead of their own interests before the end of the year (maybe even the end of November), so the designated “anti-Dean” candidate—Dick Gephardt seems like the only alternative with enough Old Left street cred, regional ties in the midwest swing states, and establishment support—can gain sufficient traction against both Dean and the novelty candidates. And if you see John Kerry, John Edwards, or Wes Clark stepping aside to back Gephardt, you’re truly kidding yourself.

My current theory on how the nomination battle will play out is explicated here.

James Joyner essentially agrees.