11/30/2011 8:11 PM
Illegitimate BCS process holds game hostage
By Dan Wetzel, Yahoo! SportsNov 29, 11:19 pm EST
The BCS formula is an exercise in nonsense.
It always has been; it's just more obvious this season, when there is a heated debate over the second-best team, the one that would meet LSU for the title.
Alabama, Oklahoma State, Virginia Tech and others are making their pitches, pointing out this strength and that argument to get a crack at the Tigers. The campaigns will only pick up this weekend.
Understand this, though: No matter what it says, the BCS is not a system designed to choose a championship matchup. It is merely a tool to stave off the inevitable playoff bowl directors fear will cut into their millions in tax-free profits, a casino-style distraction to placate the masses.
It is what it is, and until it collapses (even a four-teamer is a major, positive step), college football is stuck.
That said, if the BCS somehow survives in its current incarnation, the formula to determine 1-2 must be scrapped.
It currently consists of two-thirds human opinion polls that are ripe for political foolishness, full of oft-uneducated voters and subject to groupthink.
The other third features an average of six computer formulas, which quantitative analysts have declared mathematically unsound and their own proprietors admit are not as accurate as they could be. Five of the computer formulas are secret, even kept from the BCS, which means no one, absolutely no one, knows if they are accurate or honest.
It is a total disaster of a system. No one who cares about the game would ever invent such a thing. Just because ESPN does a fine job dressing it up each week like it's a legitimate process doesn't mean it is.
There are two obvious solutions to this problem.
1) A single public and reputable computer formula that would allow every team to know what it is dealing with from the first snap of the season. Whether such a formula could be created is the chief question. If it's possible, you might not like it but you'd at least know the rules of the game.
Or:
2) A small selection committee of five or seven people. Much like the NCAA basketball tournament, that group could meet and knowledgably assess the contenders using pre-set criteria and common statistics in a calm, rational and unhurried process that doesn't begin until the season is actually over.
The former is unlikely in a sport run by people with an obvious distaste for math. If they won't allow the current formulas to operate honestly, it doesn't seem likely they'd turn the entire process over to a computer.
The latter is potentially even more problematic. It would require actual individuals to stand up and defend the decision and, by proxy, the entire system. It's much easier for the suits to shrug their shoulders at the matchup and blame faceless "computers" or a throng of "voters."
BCS executive director Bill Hancock likes to refer to the formula as "one part science, two parts art." It's actually just a wholly formed insult to intelligence.
The computer formulas are mathematically laughable. There is plenty of debate about how accurate any ranking formula is when trying to use limited data (12 or 13 games) to sort through a diverse mass of teams. Some of the formulas include not just the 120 FBS teams but extend all the way to junior colleges.
In late October, Kenneth Massey's rankings had Arizona Western, a community college, 30th overall.
Seriously.
Then the BCS corrupts the process in a number of ways. First it asks the formulas to spit out results starting in midseason, with half the data. Those results influence the human polls, only the humans are historically slow to reassess a team unless it loses a game. The computer number may drop later, but the poll number won't. Essentially it's asking for the results of a half-finished equation, which then corrupts all the data.
Even more notably, the BCS does not allow margin of victory to be included. As such, a 70-3 victory counts the same as a 7-3 win.
The BCS' stated reason? Sportsmanship.
Not only is sportsmanship a debatable concept to begin with, it is not a mathematical property and can't be incorporated into a math equation.
The entire effort is undermined by coaches routinely running up the score anyway because human voters pay close attention to margin of victory. So in application, the BCS doesn't stop "unsportsmanlik" play, it rewards it. Besides, computer operators say they easily could put a cap on the margin to at least account for a three-touchdown triumph.
The BCS doesn't care. It prefers what Jeff Sagarin calls the "politically correct" version.
"You're asked to rank teams that play each other, that don't play long seasons, and you can't include margin of victory?" Massey said. "It does require sacrificing a bit of accuracy. It's not the best way to do it."
Massey, among others, uses margin of victory to produce what he calls a "better version" of the rankings on his website. The BCS prefers the worse version.
UC Irvine mathematics professor Hal Stern published a lengthy takedown of the BCS by in the Journal of Quantitative Analysis. He argued the laws of math may not matter to – or even make sense to – the BCS administrators. Anyone wonky enough to rank college football with math, though, does get the sanctity of numbers – and he believes the pocket-protected should take a drastic measure.
"I am advocating a boycott of the Bowl Championship Series by all quantitative analysts," Stern wrote.
"Stern's analysis was clearly right," said Bill James, the baseball statistics maven whose work was highlighted in the book and movie Moneyball and backs Stern's call for a BCS boycott. "This isn't a sincere effort to use math to find an answer at all. It's clearly an effort to use math as a cover for whatever you want to do.
"It's just nonsense math."
And that's if the nonsense is accurate. The truth is, no one knows. Five of the six computer rankers protect their formulas, claiming proprietary rights to their algorithms. They refuse to share the formula with anyone, including the BCS, which, as such, is incapable of checking on the accuracy of the rankings.
Repeat: Absolutely no one has any idea if the weekly BCS standings are true. No one. And they don't seem to care.
Only Wes Colley makes his formula public, which allows outside review. Last season, his rankings twice were found inaccurate because of improper data entry. Some of the scores he used weren't up to date.
It was only discovered because Jerry Palm of CollegeBCS.com decided to double-check Colley's rankings. This included the final standings, where a failure to enter the result of a game between FCS programs Appalachian State and Western Illinois caused a change in the final rankings of four teams in the top 20, including LSU and Boise State.
It was mere happenstance that it didn't affect the championship matchup. Of course, there's no telling if the other five were accurate.
Hancock, the BCS executive director, called it unacceptable, set up a "peer review" of result entry and asked each ranker for a synopsis of his formula. Then he acknowledged in an interview last spring that the BCS doesn't have the full formulas (if it could even comprehend the higher mathematical equations) and there remains no way to check for accuracy. They just trust the computer folks.
So those final rankings Sunday might be correct. Or they might not.
"The BCS still doesn't have any idea what these guys are doing," Palm said. "This is my longstanding gripe with the system. No one is held accountable. There is no reason why they don't use open computer formulas except they either don't understand or they don't care."
Then there are the two polls, the Harris Interactive and the coaches' poll. Each has a host of problems, the most obvious that human voters are susceptible to media coverage, political pushes to hype one team over another, historic brand names and the fear of standing out from the pack.
By doing weekly polls, stagnation also can set in. There is no reason to run BCS standings in October or November other than to gin up hype for a college football season that shouldn't need anything artificial.
By doing so, a team's strength and weakness can be set before a full body of work reveals itself. Many voters admit they don't like moving a team down unless they lose a game. But why are they up there in the first place if only one poll – the last one – matters? There should only be one poll so there isn't the current problem of the earlier, pointless polls impacting the final decision.
The coaches' poll even starts in preseason, meaning projections can influence reality. What appears to be a quality win early in the season, thus pushing a team up in the rankings, may lose its significance later. Rarely do voters take a corrective measure.
Then there is this: Coaches readily admit they watch terribly little college football – other than scouting footage of their upcoming opponent. Many don't even fill out their weekly ballot, a task left to secretaries or sports information directors.
"I don't know why we vote," South Carolina coach Steve Spurrier said. "I guess we vote because college football is still without a playoff system. I really believe most coaches do not know a whole lot about the other teams."
On Monday, the Wall Street Journal published the results of a detailed, multiyear study of coaches' voting patterns from Yale economist Matthew Kotchen and UC Santa Barbara political scientist Matthew Potoski that highlighted various biases.
"Coaches distort their rankings to reflect their own team's reputation and financial interests," the study concluded. "On average, coaches rank teams from their own athletic conference nearly a full position more favorably and boost their own team's ranking more than two full positions. Coaches also rank teams they defeated more favorably, thereby making their own team look better."
This doesn't just influence the 1-2 matchup. It plays a role in how the 1-2 matchup is contemplated, since the number of victories over top-10 or top-20 teams in the BCS standings are a reasonable way to compare resumes.
In essence, much of the supporting information to make a decision has been poisoned.
The Harris Poll, which the BCS created when The Associated Press demanded that its weekly rankings no longer have anything to do with the process, may be even worse than the coaches' vote. It consists of 115 voters, each nominated by a particular conference, many with strained ties to the current game.
There are retired refs and sports information directors. Some are former players and coaches, plus a bunch of media and college administrators. Most have lives outside of college football and, quite understandably, can't watch all the action. They have their own issues with bias and bizarre votes. They struggle to see lower-profile teams. They've proven easily swayed by campaign slogans and teams that run up the score. And occasionally they straight up forget to vote.
Of course, the BCS provides little guidance on how they should do their job, preferring to leave an organically created mess.
Does strength of schedule matter? Winning a conference championship? How about whom you beat? Or who beat you? Or by how much? Or when? Or where? Does a team's style of play count? Should voters attempt to put together the best matchup, or simply chose the two best teams?
No one knows. It's up to each person. Often the reasoning is arbitrary and senseless. Some of the buzz in the media includes a moral opposition to a rematch or that some people viewed the Alabama-LSU game as boring and thus not worth watching again.
Is a voter's predicted personal entertainment really a variable here?
No matter where you stand on who should be selected for this season's title game, the process is ridiculous.
The players and coaches deserve to settle things on the field in a playoff – which they'd prefer no matter the baseless poll numbers the BCS trumps. If not, then at least use something other an imbecilic charade of a system.
A selection committee may not be ideal, but at least it would be an accountable, professional group that is dialed into each of the candidates. Then, equally committed experts can challenge (or support) their findings. Nothing would be debated or set until the entire season is done and all resumes complete.
No one would have the advantage of being in second place before the process began. To use this season as an example, since computing and voting for the only rankings that matter won't occur until Sunday morning, then why do Virginia Tech, Oklahoma State and the others have to make the case for bumping Alabama out? There shouldn't be anyone "in." Nothing should have happened yet.
Players, coaches and fans may or may not agree with the final result. They could at least understand there was an honest attempt to solve an impossible riddle: pick two teams.
The reason this doesn't exist is because the BCS suits that claim they believe so thoroughly in the system don't have the courage to defend the decision and be the face that makes the tough choice. That's a thankless job, and who wants to fall on their sword for the BCS? They'll flock to the open bar on some bowl-game-paid Caribbean cruise, but they aren't taking the heat for this joke.
The current formula is nothing more than nonsense math and an unsound popular vote that gets polished up by television. Anyone who cares about college football should demand something better.
If we can't get a playoff, can they at least stop insulting our intelligence? Can we at least get some attempt at sanity?