When the new College Football Playoff was announced with great fanfare and a star-studded Committee was introduced, the college football world rejoiced: No more computers! And no more looking at the polls! But it's pretty clear that computers and polls are what the Committee is going to use to make its decisions, which will probably look a lot more like the BCS results than anyone wants to admit.
Let's take the polls. The AP and coaches poll are created by sportswriters and coaches. The Harris poll (used by the BCS) was a mix of former players, coaches, and media members. The College Football Playoff committee has the same makeup—current Athletic Directors mixed with ex-coaches and players. The names may be "bigger" but there's no reason to believe their choices will be more informed than the aggregate of sportswriters or coaches in the AP and USA poll.
Then there's the computers. While the six systems used by the BCS are gone, the Committee has signed up with a service that will provide a variety of statistics. They'll take those into consideration the same way the NCAA basketball selection committee uses the "Nitty-Gritty" report to select its 68 teams in March. The difference is there is no "RPI" equivalent supplied for the football committee.
But should there be? Tons of information with no way to digest it can be used to justify anything. A good power rating provides a way of digesting thousands of points of information. Case in point: Barry Alvarez describes a very basic and important kind of power rating when he talks about criteria he will use:
"In other words, let's say a team scores 50 points four games in a row, but the teams they're playing give up an average of 45," Alvarez said. "And then you've got someone who's played a very difficult schedule and they're averaging 21 points a game, but the teams they're playing only give up an average of 10 points. There's a huge difference there. That variation on both offense and defense are very telling statistics."
Hmm, let's see. Comparison of points scored to the opponent's average points yielded? And vice versa? And if we take that through multiple cycles until the numbers don't change? Then you have a margin-of-victory based power rating. Much like our Strength power rating, or Sagarin's predictor, or the SRS from pro-football-reference.com.
ESPN's article makes pains to distance what he's saying from margin of victory, which was verboten in the BCS era:
While Alvarez said he doesn't place much stock in margin of victory, he does pay close attention to a statistic on relative offensive and defensive performance.
Which of course is impossible. If you are paying "close attention" to relative offensive and defensive performance, you are ultimately paying close attention to margin of victory. You are just adjusting it for strength of schedule, which is what a power rating does.
And this is what is going to happen as the Committee becomes more versed in how to rank college football teams. There's no doubt the Committee has tons of college football expertise, but there's no indication at all that they have any expertise in ranking college football teams. The two are not nearly the same thing.
Right now their focus is on information, and deciding what information is relevant. Perhaps each member will have "pet" statistics to use. They might each use a different method to decide what "strength of schedule" means, and it's important since that's pretty much all we've heard them talk about since the committee was formed.
Maybe as the committee members think about the problem more—what constitutes quality?—then they will replicate some of the solutions that have already been used to rank teams in the past. Alvarez took the first step in replicating a Strength/Predictor/SRS power rating. Others might come up with MOV-free descriptions. They might have programmers implement some of these systems. If, say, half of the committee members do so, then there will be six computer programs that will be part of the Playoff decision process.
Add those six computer programs to the Harris/USA-like mix of coaches and football players on the decision Committee, and it will be like the BCS never went away!
Ohio State busts into the bracket: what hath the Committee wrought?
Ohio State jumps into the bracket, sure to cause controversy in the Big Twelve world where TCU—#3 last week and a 55-3 winner over the weekend—fell three spots.
The Buckeyes beat Wisconsin 59-0, which is certainly impressive. The Committee apparently was impressed with the fact that Ohio State "lost two quarterbacks"—even though one of them didn't play a down this season, which is somewhat like a team losing its quarterback to graduation or the NFL.
The decision shows a lot of shortcomings of the Committee, mainly, releasing weekly rankings. Though their final rankings showed they didn't fall into one "trap" of weekly rankings—namely, moving teams up and down instead of re-evaluating—but they did fall into another trap: teams that lose early have a better chance of having that loss "excused." Ohio State's lose to Virginia Tech was by far the worst of the eligible teams, coming at home to a 6-6 team by 14 points.
If the Committee is going to release weekly rankings, they should release point totals to reflect how close the race is; that would have reflected how close the voting between #3 and #6 was. A better solution would be to not release weekly polls at all.
In the end, TCU and Baylor fell victim to not having a championship game, and therefore, in the Committee's mind, no conference championship. They also fell victim to another problem with the Committee: a school's name means a lot.
Let's not fool ourselves: Ohio State is a "name" football school and TCU and Baylor are not. It didn't matter that TCU beat Minnesota convincingly and Ohio State won by 7 points. It didn't matter that Baylor has the best win of any of the top six teams. It didn't matter that Ohio State won the Big Ten, which was pretty atrocious this year. In the end, it mattered that Ohio State was Ohio State, and they had a huge, convincing win at the end that allowed the Committee to say the Buckeyes passed the "eye test."
Here's a question: if TCU and Oklahoma's names were switched on their schedules, would Oklahoma have been dropped from #3 to #6 after beating Iowa State 55-3? And if Oklahoma and Baylor's names were switched, would Oklahoma ever have been ranked behind a TCU team that they beat?
People hated the BCS because of the computers, but mainly because there wasn't a Playoff. Now we're faced with a very arbitrary Committee, and still have teams that might be the best in the nation left out of the Playoff.
Reblog (0) |