The BCS formula is made up of
three parts and breaks down as follows:
1.
Harris
Interactive Poll: (115 voters × 25 pts. = 2875 total possible) Each voter
ranks teams 1 through 25, and then the #1 team on that ballot receives 25
points, the #2 team receives 24 points, and so forth until the #25 team
receives 1 point. Each team’s total number of points is then divided by 2875 to
reach the Harris Poll Percentage.
2.
USA Today Poll:
(59 voters × 25 pts. = 1475 total possible) The same process described for the
Harris Poll is the process used to reach the USA Today Poll Percentage. The only difference is that each team’s
total number of points is divided by 1475.
3.
Six Computers Ranking Providers are used—Anderson & Hester,
Richard Billingsley, Colley Matrix, Kenneth Massey, Jeff Sagarin, and Peter Wolfe: After
each team’s highest and lowest computer rankings are dropped, the middle four rankings
(each having 25 pts. possible) are added and divided by 100 to reach the Computer Rankings Percentage.
4.
The BCS Standings are then reached by
averaging the percentages from 1–3 above.
You can read an adequate, though
inaccurate, BCS
explanation on ESPN.com or a more lengthy description on Wikipedia.
Most significant in the final BCS Standings is that the
computers and the human voters agreed on the ranking of only three teams in the
top twenty-five: LSU, Arkansas, and Nebraska. I highlighted these three teams in the list below. Yet the two separate human polls agreed with each other on the rankings of all twenty-five teams.
I also highlighted the teams that ended up with a lower
ranking in the Final BCS Standings based solely on the “eye test,” or the human
voters. Those teams are Oklahoma State, Kansas State, Oklahoma, Baylor, Georgia, TCU, Houston,
Texas, and Auburn. By contrast, the teams that are not highlighted ended up
with an inflated ranking in the Final BCS Standings because of the same human-voter “eye test.” It is interesting to see some of the very divergent differences in the rankings.
Final Computer
Rankings:
1.
LSU (13-0)
2.
Oklahoma State (11-1)
3.
Alabama
(11-1)
4.
Kansas State (10-2)
5.
Stanford
(11-1)
5.
Arkansas (10-2)
7.
Oklahoma (9-3)
8.
Oregon
(11-2)
9.
Boise
State (11-1)
10.
South
Carolina (10-2)
11.
Baylor (9-3)
12.
Georgia (10-3)
13.
Virginia
Tech (11-2)
14.
Wisconsin
(11-2)
15.
Michigan
(10-2)
16.
Clemson
(10-3)
17.
TCU (10-2)
18.
Houston (12-1)
19.
Texas (7-5)
19.
Nebraska (9-3)
21.
Auburn (7-5)
21.
Michigan
State (10-3)
23.
Penn
State (9-3)
24.
Southern
Miss. (11-2)
25.
West
Virginia (9-3)
Clearly any ranking system will
have its problems of “fairness.” But a computer ranking system is objective—or at
least it purports to be objective. Yet it quickly becomes evident that all the highlighted teams are either
Big XII (or future Big XII in the case of TCU) or SEC teams, with the sole
exception being Houston, which dropped one spot in the Final BCS Standings. The
computers favor these two conferences over the rest, particularly the Big XII.
But why?
While I’m not as familiar with the
SEC, the Big XII is an extremely difficult conference to compete in (but from the way this season has played out, I’m sure the same is true for the SEC).
Oklahoma State’s one loss, Kansas State’s two losses, Oklahoma’s and Baylor’s
three losses each, Texas’s five losses, and even one of TCU’s two losses all came at
the hands of Big XII teams.
The same holds true for the five Big XII teams that
did not finish in the top twenty-five. All six of Iowa State’s losses, all
seven of Texas Tech’s losses, and nine of Kansas’s ten losses were to Big XII teams. And although both teams are
headed to the SEC next season, Texas A&M and Missouri each suffered five of six and four of five losses a piece to Big XII teams.
What does this mean?
First, the
Big XII’s non-conference record this season was an amazing 27-3. The three
losses were to Georgia Tech (will play Utah in the Hyundai Sun Bowl), #6
Arkansas (will play #8 Kansas State in the AT&T Cotton Bowl), and Arizona
State (will play #7 Boise State in the MAACO Bowl Las Vegas).
Second, the Big
XII tends to “shoot itself in the foot” when it comes to having any standout
team that clearly dominates the rest of the conference. That’s how Oklahoma—holding
the #1 ranking preseason through week 5—lost at home to an unranked Texas Tech
team having a bad season (5-7 overall). That’s how Oklahoma State—starting at #9
preseason and moving to #3 by week 9— lost in Ames to an unranked Iowa State
team (6-6 overall) in double overtime.
In the end, was Oklahoma State's loss to Iowa State “worse” than Alabama's loss to LSU? Or were Oklahoma State's wins “better” than Alabama's wins? The Computer Rankings Percentage clearly indicates that the answer to these questions should be, “no, yes.” The human voters were closely divided, but the “eye test” answered these same questions, “yes, no.”
In addition to the travesty that Oklahoma State will not get a shot at the National Championship is the fact that #8 Kansas State and #7 Boise State were both passed over for at-large bids. And #18 TCU was clearly prevented from moving up to #16 to receive its Automatic non-AQ bid. For their part, at least the human polls put TCU at #15. But the problem was still in the fact that the human polls put two, three-loss teams and Michigan ahead, thus preventing a TCU appearance in the BCS.
And how about the finagling that somehow moved an idle Michigan team from #16 the week before to #13 in the final standings? In order for a team to qualify for an at-large bid it must finish the season in the top fourteen and have at least nine wins. The ESPN commentators were all shocked on Nov. 27, when they saw that Michigan was #16. What kind of message did the commentators intend to send the voters?
“Michigan is supposed to get an at-large bid, so make sure that Michigan gets moved up at least two spots next Sunday.”
TCU ended its season 10-2, played its final game on Dec. 3, and was the Mountain West Conference Champion. Michigan also finished 10-2, played its final game on Nov. 26, but did not even win its division, let alone its conference. And how was each team rewarded for its efforts? TCU remained at #18, two spots shy of an Automatic Non-AQ bid, while Michigan leaped ahead to #13, grabbing an at-large bid to play against #11 Virginia Tech, another at-large bid, in the Sugar Bowl.
How did this “Michigan magic” happen? It's rather simple, really.
Number 10 Oklahoma (CPUs: #5, USA: #11, Harris: #10) lost the de facto Big XII Championship Game to #3 Oklahoma State.
Number 13 Michigan State (CPUs: #16, USA: #9, Harris: #11) lost the inaugural Big 10 Championship Game to #15 Wisconsin.
And #14 Georgia (CPUs: #14, USA: #14, Harris: #12) lost the SEC Championship Game to #1 LSU.
Michigan State was already ranked higher in the human polls than it should have been, so it was easy for both human polls to drop Michigan State minimally to #13 (USA: -4 spots, Harris: -2 spots) and allow the Computers to do the dirty work of dropping Michigan State to #21 (CPUs: -5 spots), leaving Michigan State at #17 in the BCS. Michigan State dropped four spots because it played in the Big 10 Conference Championship Game against the #15 team and lost.
Georgia, too, was easy. The human polls knew what to do and both somehow put Georgia at #18 (USA: -4 spots, Harris: -6 spots). As to the Computers, they actually moved Georgia up to #12 (CPUs: +2), leaving Georgia at #16 in the BCS. Georgia dropped two spots because it played in the SEC Conference Championship Game against the #1 team and lost.
Oklahoma, however, was not as easy to move behind the idling Michigan team as were the other two teams. This is because the Computers take into account strength of schedule and reject the “eye test” altogether. The Computers only dropped Oklahoma to #7 (CPUs: -2). Yet both the human polls somehow felt the need to drop Oklahoma to #19 (USA: -8, Harris: -9). This move by the human voters was by far the most ludicrous part of the Final BCS Standings. But tinkering with the system worked, and it left Oklahoma at #14 in the BCS. Oklahoma dropped four spots because it played in the Big XII de facto Conference Championship Game against the #3 team and lost.
What is the takeaway message? Both human polls ranking each of the top twenty-five teams in the exact same order as each other is no mere coincidence. Take the human “eye test” out of College Football Rankings. While one eye might be watching the games, the other eye is always watching the money.
 |
| Final Harris Poll Rankings, Dec. 4, 2011 |
 |
| Final USA Today Poll Rankings, Dec. 4, 2011 |
 |
| Final BCS Standings, Dec. 4, 2011 |
