Monday, December 05, 2011

Why am I supposed to be furious?

According to every columnist-type person in the universe, I should be outraged right now at the idea of a rematch and the stupidity of the system and everything else related in any way to the BCS. RABBLE RABBLE SEC UNFAIR RABBLE WIN CONFERENCE RABBLE.

So ... the outrage. I don't have it. LSU and Alabama are probably the two best teams in the country and will play for the national title. This is not the worst thing ever; it isn't even a bad thing. My understanding is that the two best teams are supposed to play for the national championship. That's about to happen. Hooray?

I made this comment the other day ...
The whole thing is simple (kind of) but gets complicated by a million arguments that miss the forest for the trees.
... but didn't really expand on it a whole lot. AFAIK, the entire purpose of the BCS is to match the two best teams in the country, and that's good (at least in theory). Everything else is a byproduct that's really no better or worse than the old bowl system (not significantly, anyway). I've argued with people a lot over the last few years about that particular premise. The weird thing is that those arguments were always about Boise State. My side: I don't care at all about strength of schedule or number of wins over ranked teams or conference strength or blah blah blah. What I care about is which are the two best teams. If Boise State is/was one of the two best teams in a given year, skewing the poll because of some relevant-but-not-all-telling number is stupid. Vote for the best teams.

The problem (for lack of a better word) is trying to figure out how to define "best." Everybody has a slightly different formula that typically gets closer to entirely resume-based by the end of the season. I mean, performance is what matters, right? But there are extremely limited data points because of schedule variations, small sample sizes, etc. Houston/Boise State/whoever can go 12-0 and not be top-10 good. Houston/Boise State/whoever can also go 12-0 and be absolutely dominant. The record shows no difference. Things like total yards and scoring margin and whatnot offer some help; how much help is hard to determine since you have to do a lot of schedule-related adjusting to figure out exactly what the numbers mean. There are almost zero direct comparison points available, and a lot of the ones that exist are too skewed (because of site, injuries, circular "Team A beat Team B beat Team C beat Team A" logic, etc.) to take at face value. There are infinite variables that can't be stripped out and can't be mathematicized the way the BCS tries. What that means is that every ranking system is some sort of fluid combination of randomly timed observations, preferable stats and a simplified wins-times-strength-of-schedule formula. And then it changes to fit the argument.

I've realized recently that there's a fundamental difference of opinion about that. To me, the championship game should feature the two best teams. To other people, it should feature the two most deserving teams, which might be easier to define in a formulaic sense but is really hard to define in a fairness sense. To other people, it should feature whatever particular combination seems compelling that year, which is super subjective.

All the arguments I've seen against an LSU-Alabama title game revert to the same stupid premise: We've already seen that game once and therefore shouldn't see it again in the title game. The problem with that: It ignores the possibility that the two best teams in the country are in the same conference or division. This would be an appropriate spot to reiterate the two best teams are (in theory) supposed to play for the championship.

The one argument I really haven't seen (at least not specifically) that has some validity is the one that says it's impossible to know with 100 percent certainty that LSU and Bama are the two best teams. Alabama didn't play Iowa State or Oklahoma or Baylor and Oklahoma State didn't play Penn State or Arkansas; again, the comparison points are not just limited but nonexistent. Maybe Oklahoma State is the best team in the country. We'll never know for sure. But if the consensus is that LSU is the best team -- which seems to be commonly accepted -- Alabama has to be right there too. We know for a fact that LSU and Alabama are almost dead equal; there are infinite scenarios in which one play turns out slightly differently in that game and Alabama wins, thus becoming the unanimous No. 1 team and an indisputable lock for the national title game. The difference between a win and a loss in that game was nothing.

There's part of me that wants to see Oklahoma State-LSU for both the entertainment value and the give-me-another-data-point value. Let me put it a better way: If Oklahoma State is roughly equal to Alabama and we've already seen LSU beat Alabama in Tuscaloosa, why not give No. 2(b) a shot so we can see LSU against all legitimate challengers? To be clear, I wouldn't have an issue with that. The Oklahoma State-Oklahoma game was enough for me to believe that Oklahoma State might be in Alabama's class and therefore LSU's class. The obvious problem with that scenario is that it's setting up a de facto playoff solely for LSU. There'd be an absolute and unquestioned national champion if LSU were to win that game since that'd be a ridiculous fourth win over a top-six team, but an Oklahoma State win wouldn't necessarily mean anything definitive in that scenario other than Oklahoma State > LSU. The reason it wouldn't necessarily mean anything: We'd just have taken one particular team of similar quality to Alabama to create a new scenario since we'd already seen the most obvious alternative. But couldn't Stanford/Boise State/whoever be just as good as Oklahoma State? The lack-of-data-points issue can't be resolved by putting LSU and Oklahoma State on a field and saying, "yup, this settles it." Nay.

In the absence of a larger (non-two-team) playoff, a line has to be drawn below the consensus second-best team. That's it. The hypothetically amazing idea of seeking out all possible relevant matchups is ridiculously implausible given the bowl format. That's unfortunate for the people who believe Oklahoma State is the second-best team this year. It's a ranking-based, Republic-style system. It is what it is.

Speaking of which, there's a hilarious meme that gets rehashed every year at this time by Coach Pissed Off of Snubbed University regarding "the system" and how it's so unfair and the computers make no sense and yadda yadda yadda. We get it: You're old and cranky and don't know what Twitter is. Chris Petersen is this year's That Guy.

"The computers" are not the problem. They're valuable because they spit out hypothetically unbiased numbers based entirely (allegedly) on wins, losses and schedule strength, even if those numbers sometimes make no sense because the people in charge of the system don't allow a whole bunch of relevant information; that's probably an argument for another time. BTW, I want to punch every coach who says "I don't understand how can computers pick a national champion" the same way my great-grandma would say "I don't understand how a computer can get mail." Information goes in and information comes out. It's not freakin' Skynet putting money down on Oregon and then manipulating the results.

"The system" is not the problem. The system was designed to be an improvement on the old every-bowl-pick-some-teams days that led to a 1-2 matchup about twice a decade and a split national championship about equally often. The BCS produces a 1-2 game every year and a bunch of other bowl games that are basically equal to (or maybe better* than) what we had before.

The concept of the system is the problem: Picking two and only two teams (best, most deserving, whatever) based on the amount of information that can be gleaned from one season is a totally unreasonable task. I read a pro-playoff piece once that pointed out that the college football season provides less data/information than any other sport yet is the only season that doesn't end in a playoff of some sort. Think about that.

A playoff is so obvious that I don't even need to argue for it. Preaching, choir, etcetera. But there isn't one. There's a two-team game, and LSU and Alabama are in it because a large majority of people believe they're the two best teams. That's what's supposed to happen. There's an important distinction between the system failing and the system doing exactly what it was intended to do but failing to produce the publicly preferred outcome, which was essentially an everybody-gets-a-shot-at-LSU playoff.


Post a Comment

Powered by Blogger.