Posted: 10 December 2008 Updated: 22 May 2010: fixed some typos, added some updates, and modified the figures.
As always, this is my opinion. If you wish to communicate with me regarding this topic, use cdoswell # earthlink.net - either click on the email link or cut and paste the email address, replacing " # " with "@". Your correspondence with me, along with my response, might be added to this page as a result - if you're not willing to make your comments public, either don't send them in the first place or convince me I shouldn't make them part of this public discourse.
Recent events in college football have exemplified several ways by which I see that an argument can become worthless. I'll be addressing what I believe to make an argument meaningful at the same time. Although the major part of this is going to be about the current controversies regarding college football, it really is intended to be more widely applicable than this narrow context.
This is one of the most common problems with an argument. Imagine a debate about theories of the Universe between Albert Einstein and a congenital idiot. Even assuming that these two folks would meet and choose to have a discussion, it seems pretty evident that such a debate would necessarily be rather one-sided. It's hard for me to imagine what such a debate would be like, given the overwhelming lack of common ground that the two sides would share. It's axiomatic in debates that there must be at least some common ground upon which a logical argument can be constructed. Without any common ground, the most you can do is to agree to disagree.
Given what I see in many of the so-called debates about the current situation in college football, it's pretty tough to imagine such a debate being meaningful if you consider your opponents in the discussion to be morons. This doesn't mean that you can't disagree about the interpretation of the facts - far from it. But if you believe that a person's capability for rational thought is entirely contingent on agreeing with you about virtually everything, then you're manifestly disrespecting your opponent. It amounts to a form of closed-mindedness, and an absolutely closed mind in any argument renders the argument meaningless.
Let me state right at the outset that I'm a fan of University of Oklahoma (OU) football, and have been since I matriculated here at OU as a new graduate student. I love nothing more than seeing University of Texas (UT) football fans in anguish over their football fortunes. It adds considerably to my pleasure over OU's advancement to the BCS National Championship Game with the University of Florida (UF).* An important reason for the sweetness of the current situation is that I know in my own heart that it could just as easily be OU that was "left out" of the running for a national championship! Listening to futile weeping and gnashing of teeth by UT fans is pure joy!
There were meaningful arguments in favor of all three teams tied for the top of the Big-12 South Division - OU, UT, and Texas Tech. University (TT) - and the tie-breaker came down to which team was getting the most positive national attention at the end of the year. The schedule definitely favored OU in this respect, which had to play three very challenging games at the end of the year, whereas the opponents UT played at the end of the year were considerably less challenging. Bad luck for UT and good luck for OU, which went out and won those last three games by substantial margins. Is this necessarily fair? No. But many things in life aren't fair. Had the shoe been on the other foot and UT had advanced to the Big-12 Championship game instead of OU, I certainly would be unhappy about that outcome, but - I wouldn't thereby consider all UT fans to be brainless bozos for believing their team deserved to be there. The problem is that the BCS inevitably is unfair to someone every year, not owing to a shortage of brainpower amongst OU fans. I respect and understand the fervor of UT fans, but don't believe they're justified in seeing arguments in favor of OU playing in the BCS National Championship game as entirely without basis. All the arguments in favor of anyone's favorite team being in this game are equally valid - the tie-breaking choice inevitably was going to leave out one or more deserving teams.
I happen to be friends with quite a few UT fans, and I think we can enjoy the good-natured ribbing about the outcome of the season that flows from one team winning and the other losing, without losing any respect for each other. Do I take pleasure in the unhappiness of UT fans? You bet! And when the shoe is on the other foot, as it inevitably is from time to time, I'm sure they enjoy my discomfort. So what? They're still my friends!! The whole thing about these never ending arguments is that they're fun! They aren't life and death issues. There's no need to demonize or dehumanize opposing fans. Agreeing with me is not a criterion for my respect or my friendship, and my UT fan friends are equally rational about the situation, despite their foolish commitment to the wrong team! <);-)
A significant fraction of the content of message boards on this subject is dominated by abusive, mean-spirited remarks about the fans of the opposing team, to the point where whole states are being condemned as slimeballs and morons. When someone makes such blanket statements about the fans of another team, they often seem to be forgetting that their own fans may be guilty of similar behavior. I make no denials that some OU fans also participate in abusive "arguments" in these message boards. There's no more excuse for that than comparable remarks from UT fans (or any other team).
Another essential element about a meaningful argument is that all parties to the argument must agree on how to define the terms they'll be using. This is related to the issue of "common ground" that I mentioned above. If the debaters disagree about the terms they're using, then it's very unlikely they can come to a meaningful agreement about the point in question. When it's logically impossible to come to agreement about the point, then the debate is meaningless. The whole point of a logical argument is to change someone's mind. In cases where the disagreement begins with definitions, the rest of the debate is pointless unless that can be settled and agreed-upon definitions become the basis for further debate. If one side (or both sides) in a disagreement over defining terms is closed, the argument is rendered meaningless.
Consider the ongoing argument over who should get the Heisman trophy for NCAA football in 2008. The general consensus is that the three top candidates are Colt McCoy (UT), Sam Bradford (OU), and Tim Tebow (UF). All three fan bases support their candidate and offer various arguments in favor of their team's candidate. Of course, there are other candidates besides this trio, each of whom have valid arguments in their favor, but they're unlikely to break into the top three, much less win the Heisman trophy - fair or not. Various interpretations of the factual information can be used to buttress the candidacy of any candidate. The Big-12 apparently has better offenses and weaker defenses than the SEC, so Tebow is clearly the top candidate because of the stronger defenses he had to overcome to win. Apparently, OU and UF have better skill players on offense than UT, so Colt McCoy is better because he has to do more with less for his team to win than either Bradford or Tebow. Bradford's statistics apparently are off the scale when it comes to touchdown passes, interceptions, etc., so his numbers make him better than McCoy or Tebow.**
The problem with these arguments is that although they have varying degrees of validity (I'll get to that later), they're all logically valid interpretations of the facts to favor one candidate or another, whereas the notion of the Heisman Trophy recipient being the nation's "best collegiate football player" is very ill-defined. Traditionally, it goes to offensive rather than defensive players, which is not a necessary limitation on deciding which football player is the "best in the nation". Why couldn't the best player be on defense? How do we define the "best collegiate football player"? By statistics? Sort of. By the quality of the team around the candidate? Well, sort of. By consistent performance through the season? Sort of. By career accomplishments? Sort of. By potential to make it in the NFL? Sort of. It seems to me that every year, the combination of ingredients that go into picking the Heisman trophy winner varies, depending on who's out there for folks to consider. Minor college players are apparently out of the running altogether - perhaps because of the perceived caliber of the opposing teams, despite the established fact that many Heisman winners fail to succeed in the NFL, whereas many minor college players do succeed in the NFL.
Best by comparison with whom? Although the notion of "best collegiate football player" leaves the issue pretty much wide open, the tradition always has been to select quarterbacks or running backs from a limited set of "big-time" schools contending for the mythical national championship in Division 1-A (FBS). If it must be an offensive player (for some as-yet unspecified reason), why not a tight end? Why not an offensive lineman? Why not a blocking back? We can have all the debates we want, but without a clear definition agreed upon by all participants, the arguments are all essentially meaningless. They generate a lot of sound and fury, and may even be fun, but they signify nothing, in the end. Someone gets an award and everyone else is in last place, fair or not.
When a debate involves quantitative comparisons, it's important that the comparisons be appropriate and robust. This issue is frequently described as "comparing apples and oranges" - the essential point is rather more complicated than that, but the metaphor captures at least part of the issue. There are times when comparing apples and oranges is entirely logical, as in a quantitative comparison of their nutritional value, for instance. But I digress.
In college football, the comparison typically involves one or more of the quantitative aspects of the game (scores, rushing offense, passing defense, turnover margin, etc.). Team A beats Team B, Team C then beats Team A, and Team B then beats Team C. Which team is best? If the games don't produce a clear-cut winner, statistical comparisons of various sorts can be used to buttress arguments in favor of one team over another. This was precisely the problem in the Big-12 South Division this year. Not only does this set of game outcomes leave the issue of which team is best in doubt, but comparing the specific outcomes is ultimately flawed, for numerous reasons:
In meteorology, we've learned that a single "deterministic" forecast is not always the best way to make a prediction. Rather, we meteorologists have concluded that we're better off with running multiple solutions, under varying initial conditions and with different models, rather than relying on any single run. In college football, even "head-to-head" outcomes don't always determine which team is the "better" of the two teams on the field. How do we define which team is "best" (see point #2)? A single-elimination tournament doesn't always wind up with the best team as its champion. Seedings and the factors listed above all enter into the way such a tournament plays out. The equivalent of an ensemble would be to vary the player lineups and seedings randomly and play hundreds of games amongst all the teams, with the final decision about which team is best determined by the average of the ensemble of outcomes. Such an ensemble is completely impractical in NCAA football, but if done, it would make a much stronger and robust statement about overall team quality.
To the extent that we can look to head-to-head outcomes as deterministic, we can bask in the joy of winning or hang our heads in disappointment over losing, but such outcomes really are not deterministic. In many cases where rematches occur (as in bowl games or conference championships), they often go the other way from the first game between a pair of relatively evenly-matched teams.
Another meaningless argument is that carried on by the Big-12 and Southeastern Conference (SEC) fans, especially in light of the upcoming bowl games, highlighted by the BCS National Championship between OU and UF, but including the TT-Ole Miss Fiesta Bowl (It's a shame UT couldn't have played Alabama in a bowl game.). The Big-12 had a lot of offensive production this year and not a lot of defense, apparently. Since the adage is that "defense wins championships" it would seem that the SEC should be favored in both of those head-to-head SEC/Big-12 matchups. Just how poor are the Big-12 defenses? The statistics might be somewhat skewed by the amount of offensive firepower in the Big-12, including OU, UT, TT, Oklahoma State, Missouri, Kansas, and at the end of the season, Nebraska. Therefore, the Big-12 defensive statistics might be somewhat biased by the overwhelming impact of considerable offensive capability. And having played against those high-powered offenses might be advantageous when confronting a team with an excellent offense (such as UF). If the Big-12 teams have moderately good defenses (despite poor stats), they could be surprisingly good with a month to prepare for their bowl opponents. The bowl game outcomes might well confer bragging rights for the winners, but the nature of the games has yet to be determined. A direct interpretation of the data to imply vastly inferior Big-12 defenses might be an exaggeration. In 2000, the BCS National Championship game between OU and Florida State University (FSU) was supposed to be an offensive explosion, which would favor FSU, but turned out to be a defensive game, with OU's defense playing inspired and completely shutting down FSU.
Note that I'm not saying that game outcomes are completely random and without any meaning whatsoever about overall team quality. But when comparing relatively evenly-matched teams, past outcomes and piles of statistics are not completely reliable indicators of which team is 'best'. In fact, when we consider comparing teams, deciding which is 'best' can be difficult, not only for all the reasons I've listed, but because there's no completely robust definition of 'bestness' (save perhaps my impractical notion of an ensemble of outcomes under varying conditions). It's a concept we love to argue about, but without such a definition, the arguments ultimately are pointless.
If all the factors that we know about and measure were completely determinative, then it simply wouldn't be necessary to play the games. Predicting outcomes would be an exact science and all of the fun would be taken out of college football. If we can't predict the outcomes with perfect accuracy, and no one can, then there's always uncertainty about the outcome. Upsets are a symptom of this. OU 'upset' FSU for the 2000 National Championship. In 2008, OU was 'upset' by UT, then UT was 'upset' by TT, and then TT was 'upset' by OU. The arguments that ensue from this set of outcomes can be fun, but aren't very meaningful.
Assuming UF will be favored, will OU 'upset' UF in the BCS National Championship Game? I don't think anyone can predict that with absolute certainty, but at this point, anyone who thinks it's a foregone conclusion doesn't understand the realities of college football and has forgotten a long history of upsets in big games.* In the end, even the game's outcome isn't a perfectly accurate indicator of 'bestness' of the teams. One team will win, and its fans will celebrate - the other team's fans will mourn and some may offer excuses. But there's always next year ...
Another common problem is seeing what you want to see as opposed to seeing what's actually there. There's nothing that prevents any of us from having a prejudiced view of the situation, and that's something that anyone wishing to have a meaningful argument needs to guard against. This is another form of misinterpretation of the facts, but is associated with prejudices. If you recognize you have a prejudice, you have to accept that someone else who has an opposing prejudice could evaluate the same evidence differently. If the results are not so unambiguously one way or another, at the very least, you should understand why someone might interpret the facts differently. It might be that you couldn't sway your opponent, but you wouldn't be so inclined to disrespect that opposing view. And you might learn something useful from having to defend your interpretation from someone with a different version. A closed mind in an argument over interpretation of the evidence renders the argument meaningless.
Consider the (generally unsubstantiated) statements that have surfaced by some of the UT fans in sports message boards that OU has been "running up the score" late in games when the outcome was no longer in doubt. These fans have claimed (without evidence) that UT has steadfastly and consistently avoided doing that, whereas OU has been doing so in a cynical way to influence the poll ranking voters. Let's look at the numbers and compare the two teams in terms of their scores during the season - first, the UT season results:
UT season scores
and next, the OU season results:
OU season scores
The average margin of victory for UT is 25.33 points, while for OU, that average margin of victory is 29.69 points - less than a touchdown separates the two teams' winning margins. That's not very compelling evidence of OU cynicism and UT idealism. So, what might "running up the score" actually mean? In the absence of any standard definition, I'll propose my own: assume that this can be measured by the number of points scored in the 4th quarter of games when your lead at the end of three quarters is 21 or more points (three touchdowns or more). If I exclude from consideration any losses (which are self-evidently not trying to run up the score) and end-of-the-3rd quarter margins of 20 points or less (when the outcome of the game could reasonably be considered still to be in some doubt at the start of the 4th quarter), the average number of 4th quarter points scored by UT (in the seven games when leading at the end of three quarters by 21 points or more) is 7.29 - just over one touchdown. If I perform the exact same calculation for OU, their average 4th quarter points (in the nine games when leading at the end of three quarters by 21 points or more) is 5.78 - somewhat less than one touchdown. I would infer from this calculation that OU's propensity for running up the score on hapless opponents at the end of the game is actually somewhat less than that of UT's this year, although I believe the difference between them is pretty much without a lot of statistical significance. By this measure, they're virtually identical in this respect. Consider the games when leading at the end of the 3rd quarter, but by less than 21 points: UT's 4th quarter points scored in those games (UT had four such) was 12.50, whereas in OU's three games when leading by less than 21 points at the end of the 3rd quarter, they scored an average of 11.33 points in the 4th quarter. Even with lesser leads, OU still scored fewer 4th quarter points than UT. In the first half of OU's games, they scored 456 points, whereas in the second half, they scored 246. In the first half for UT, they scored 277 points, whereas in the second half they scored 245. The big difference between the teams was OU's complete offensive dominance in the first two quarters of most of their games. Their blowouts weren't the result of "running up the score" late in the games. Anyone watching the games should know this already!
Frankly, I find this whole discussion about OU running up the score to be without much meaning. Perhaps it does happen, for various reasons, but I agree with TT Coach Mike Leach - this whole topic is basically a crock. If you don't like what our offense is doing when it runs the offense, then your defense should stop it. If your defense can't stop it, that's not our fault. Should we just run three-and-outs for the rest of the game? Might that not be interpreted as disrespectful, as well? The last touchdown scored by OU in the Big-12 championship game could be seen by opposing fans as running up the score, except that OU was going for a record fifth consecutive game scoring 60+ points, and it was a running play, not a pass. However, it could also be a case of Coach Bob Stoops wanting that record as a reward for his players. No fan has a direct link to the thinking of a coach, and fans will tend to think what they want, no matter what the coach says - but I believe there's essentially no objective evidence that the 2008 Sooners made it a habit to run up their point totals on hapless opponents.
There have been instances in his UT career where Coach Mack Brown could have been (and was) accused of running up the score for "style points" at the end of the season. He didn't exactly shut down the offense in the 4th quarter against Texas A&M in 2008, despite leading by two touchdowns at the end of the 3rd quarter. Some fans seem to have selective vision or selective memories when it comes to this issue. When they get jumped by another team in the polls, this topic seems to come up frequently as an excuse. This year, OU scored the vast majority of its points in the first half, so the inflated scores come from that, not by piling on points during "garbage time" of blowouts in the 4th quarter. I can supply a detailed spreadsheet calculation for the score by quarters for UT and OU (and UF, too) if anyone wishes it - just email me a request.
I'm willing to bet that someone (likely a UT fan) will accuse me of "lying with statistics" here. Fine. Tell me how you define "running up the score" and show me the evidence associated with that definition. Then we can have a meaningful discussion about your definition and how to interpret your evidence.
* OU lost the national championship game to Florida 24-14 in a tough struggle dominated by defense. So on that basis, SEC defenses generally were tougher than those in the Big-12.
** Sam Bradford won the Heisman trophy, and was the #1 player selected in the 2010 NFL draft.