A Simple Solution
Let’s be honest—it ain’t rocket science. There could be a playoff in Division I-A football, and it wouldn’t destroy all the bowl games. In fact, it could actually restore some tradition. But it won’t happen anytime soon. (It’s contractually impossible before 2010.) And while university presidents should be commended for protecting the minor bowls (which benefit many teams and communities) and for resisting the monetary temptations of an “NFL-style playoff” (a phrase they like to use), their annual game of “BCS tweaking” makes them look nothing short of foolish.
To be legit, a playoff must be open to all Division I-A schools. To be adopted, a playoff must protect the current bowl system and allow major bowls and conferences to keep their elevated status. In other words, avoid undue change. (Some want change that would result in annual Big10-Pac 10 matchups in the Rose Bowl—but that’s really just changing back.)
There is a relatively painless way to achieve these objectives. While you may have seen similar suggestions, the proposal below combines a couple distinct features into a unique approach. First, the specifics:
Quarterfinal Games (a.k.a. BCS Bowls)
Although not required, the likely conference affiliations would be:
ROSE = Pac 10 champion vs. Big 10 champion
FIESTA = Big 12 champion vs. At Large team
SUGAR = SEC champion vs. At Large team
ORANGE = ACC or Big East champion vs. At Large team
Play-In Games (optional)
A play-in game allows two teams to compete for a BCS bowl spot. If needed, play-in games would be held in early December after the regular season is completed, preferably at home sites to reduce travel.
Semifinal games would occur on or about the second Saturday in January (negotiable), with home sites recommended to reduce travel. The championship would be held the following week, presumably at one of the four BCS bowls on a rotating basis (also negotiable).
Six conferences have received automatic invitations since 1998. Conferences receiving automatic BCS bowl invitations are responsible for designating their representatives. Automatic invitations can be reviewed (and revoked) by the BCS.
All conferences receiving automatic invitations are limited to one At Large team. (No conference may have more than two representatives in the BCS bowls—current rule). With six Automatic Qualifiers, the number of At Large teams can range from two to four, as follows:
# of Teams Destinations and/or Pairing Configurations
2 Both At Large teams are sent directly to BCS bowls.
3 One team is sent directly to a BCS bowl while two compete in a play-in game.
4 All four At Large teams compete in play-in games.
The ever-changing BCS Ranking System—combining variables such as computer ratings and human polls—is conceptually good but bad in practice. By relying entirely on a pre-set ranking system, the BCS is locked into results yet open to outside influences. (And as we have seen, poll participants can influence results.) However, I still recommend the BCS maintain some kind of ranking system, so long as it is used in conjunction with a…
The sole responsibility of a Selection Committee would be to identify At Large teams and determine their fate. The format described above offers a range of flexibility. Consider the following examples:
1) In 2004 there were five undefeated Division I-A teams. Two of these (#6 Utah and #9 Boise State) were not from BCS conferences and were outranked by other At Large teams (#4 Texas and #5 California). A Selection Committee could put these four teams in two play-in games, with the winners earning BCS bowl invitations.
2) In 2003 Oklahoma lost their conference championship game yet remained #1 in the final BCS ranking. A team that lost their league’s title match should not be asked to compete in a Play-In game (and usually the loss takes them out of consideration), but in this case a Selection Committee could choose to send the Sooners directly to a BCS bowl.
Believe it or not, once a Selection Committee identifies At Large teams, all pairings (including bowl assignments) can be made automatically. The process can be found below under Appendix A: Pairing Guidelines.
The two most significant features in this proposal are the Play-In Game and the Selection Committee. Because play-in games are optional, a Selection Committee has great flexibility in deciding when to use them. Since a Selection Committee is not bound by the BCS ranking, it can insure that all deserving teams get an opportunity to compete for the National Championship.
Oddly enough, a Selection Committee is not a good method to use when picking two teams for a title game or selecting At Large teams to fill remaining BCS bowl slots. The latter decision has serious monetary ramifications but nothing whatsoever to do with finding a National Champion. And the hope of finding exactly two title contenders has always been the flaw of the BCS. Sometimes the choice is obvious to all, but usually the process is no better than shoving a square peg into a round hole—and no computer, committee, or collection of voters can make it right. Under my proposal, a Selection Committee merely fills in the final spots of a small playoff.
This approach only requires three additional games (Semifinals and Championship), with an option for two others (Play-In Games). The BCS bowls matter every year, and even get to host a second (championship) bowl every fourth year. Schools from outside BCS conferences get an actual shot at the National Championship, but (thanks to Play-In Games) the BCS bowls are not forced to invite them just to make that possible. Meanwhile, minor bowls remain unaffected.
Perhaps more importantly, this approach restores tradition. The Rose Bowl’s agreement between the Big 10 and Pac 10 lasted over half a century, but under the BCS such a pairing only occurs about half the time. Conference affiliations can only be fully restored by ending the BCS and returning to the old system, or by adopting the playoff format described above. To be honest, I’m not a strong playoff proponent and would prefer returning to the old system rather than continue to watch the BCS goose itself. Yet a few small changes could not only restore tradition but even create an all-inclusive Division I-A playoff. Knowing that, I feel obligated to say something.
It’s hard to predict whether Division I-A will ever have a playoff, but for all practical purposes the approach described here is the only realistic option. So if—years from now—there’s a college football playoff and the format seems strangely familiar, you can say you saw it here first.
Vincent Ellerby, February 2005
I welcome all questions or comments. Feel free to contact me at firstname.lastname@example.org.
PS – There are a couple more sections to follow. The first is strictly for those who don’t mind tedious details, as it describes a procedure for how playoff pairings can be done automatically. Boring, but basically straightforward. After that is a brief commentary that deconstructs other playoff plans. Although the solution I describe is simple, it is not necessarily intuitive.
The Selection Committee designates all At Large teams. Should one or more play-in game be required, the committee specifies who will play, and where.
If teams compete in play-in games, the committee has two options. They can:
1) Declare in advance that a play-in game feeds into a specific BCS bowl, or
2) Let the final BCS ranking set bowl pairings automatically once the winners are known.
The second option is easier to implement, but both are acceptable.
The Rose Bowl pairing is automatic (Big 10 vs. Pac 10). The other three bowls (Orange, Sugar, and Fiesta) receive balanced pairings to the maximum extent possible while avoiding undesirable matchups, referred to here as Negative Contingencies. Pairing the highest and lowest ranked teams together is preferable, but strict adherence to that approach can result in contests between conference members or the rematch of a game just played. Negative Contingencies are considered in order, as follows:
1) A combined BCS ranking of 4 or less (#1 vs. #2 or #1 vs. #3).
rematch of a season-ending conference game for either team.
(Examples: Ohio State-Michigan, SEC Championship Game, etc.)
3) A rematch of a regular season game between conference members.
rematch of a non-conference season ending game for either team.
(Examples: Florida-Florida State, Georgia-Georgia Tech)
5) Conference members that did not meet during the regular season.
6) A combined BCS ranking of 5 (#1 vs. #4 or #2 vs. #3).
The above list is based on subjective criteria. Item #1 should definitely remain, but beyond that opinions may differ on what to include (and in what order). Needless to say, the list must be evaluated and set in advance. (Side Note: Even though the Rose Bowl pairing is considered to be automatic, it might be advisable to consider making an exception in the case of Item #1.)
Only three BCS bowls need to be paired, and each hosts a specific conference champion. (The Orange Bowl selects either the ACC or Big East champion, leaving the other to be listed with At Large teams.) The three host champions and the At Large teams are listed separately in rank order. Host champions are designated A, B, & C, while At Large teams are listed as X, Y, & Z.
Combination # Bowl Pairings
1 A-Z / B-Y / C-X
2 A-Z / B-X / C-Y
3 A-Y / B-Z / C-X
4 A-Y / B-X / C-Z
5 A-X / B-Z / C-Y
6 A-X / B-Y / C-Z
Once all Negative Contingencies have been applied to the above Pairing Table, the highest combination remaining is selected. If it should happen that a Negative Contingency eliminates all remaining combinations, the last Negative Contingency is allowed to stand and the best combination still available is used.
The four BCS bowl winners proceed to semifinal games. Home sites are recommended to reduce travel. Pairings are set automatically, with the two highest ranked BCS bowl winners hosting games except that only conference winners (or independents) are eligible. (For example, in 2003 Oklahoma was ranked #1 but did not win the Big 12 title and thus could not host a semifinal game.) The higher ranked home team would be paired against the lower ranked visiting team. Unlike BCS bowls, Negative Contingencies do not apply to semifinal games.
Although not required, it is assumed that the championship game will continue to be held at one of the four BCS bowls on a rotating basis.
The playoff format I described previously was derived through a long process of elimination. Therefore, it seems appropriate that I should revisit the topic and explain why other options fail. (This also gives me an opportunity to vent my spleen a bit.)
A Playoff Before the Bowl Games
“So your team lost in the playoff. Still gonna go to that bowl game?”
Let’s be honest, this approach really WOULD kill the current bowl system. I find it amusing and/or ironic that people who appear the most eager for a college football playoff can sometimes offer up good examples of why we shouldn’t have one. But be fair—while this may be a bad idea, it doesn’t prove that playoffs are impossible.
A Playoff Using the Bowl Games
“So you plan to travel to Mobile, Maui, Memphis, and Miami?”
Really, it’s not just travel requirements that doom this approach. The major bowls prefer being scheduled close to the New Year holiday, and for them to willingly line up in playoff brackets means abandoning prime dates and conference affiliations—not to mention doing battle over the perceived pecking order. My suggestion lets the BCS bowls keep their calendar spots, establish conference affiliations, and remain equally important. (They also get that championship game every fourth January.)
A “Plus-One” Championship
This seems to be the “playoff du jour” and as such is the most insidious. On the surface it appears to solve problems, but a closer examination reveals flaws.
The basic premise is that two teams meet in a championship game played a week or two after the BCS bowl games. Specifics are not clear, but only two approaches are possible:
#1) Teams play in BCS bowls according to conference affiliations, or
#2) Four teams meet in two “semifinal” BCS bowls.
The 2004 season serves as a good example of why neither approach works.
In the first option, renewed conference ties make the Rose Bowl pairing automatic but cause other bowl selections to get complicated. (For example, the top ranked team could seek a weaker opponent.) And in 2004, there were three undefeated teams from major conferences (Southern Cal, Oklahoma, and Auburn) that would have gone to three separate bowls (Rose, Fiesta, and Sugar). If all three win, how does the “Plus One” championship solve anything?
In the second option, four teams play in two “semifinal” BCS bowls with the winners proceeding to the championship. To accomplish this, one bowl gets the title game and two others host preliminary bouts. How that gets sorted out is beyond me—and the Rose Bowl’s chance for a Big10-Pac 10 pairing gets worse no matter what. Choosing the four teams has its own problems. In 2004 there were three clear choices, but what about the fourth? Texas edged out California for the #4 spot thanks to some late vote shifting, but what about unbeaten Utah? Or Boise State?
Let’s face facts. There are five major conferences (sorry Big East), and because of scheduling limitations it’s possible for each to have an undefeated team. (The Big 10 and Pac 10 could theoretically have two, but let’s not go there.) And you never know when a team from a lesser conference will stay unbeaten. So under existing circumstances, it’s not realistic for a playoff format to include less than eight teams—which is what you get with four BCS bowls.
“A Playoff Would Diminish the Regular Season”
Okay, so that’s not really a playoff format. But it IS a statement that’s been bandied about by those who oppose a playoff. While I understand resistance to playoff proposals, I can only respect valid arguments—and that ain’t one of ‘em.
The statement may stem from basketball and its season-ending tournament. In theory, Podunk University’s hoop squad can go on a hot streak and win the National Championship—even after a mediocre year in a pissant conference. The sentiment is that the tournament seriously diminishes the regular season, and in this case I completely agree. A better resolution would involve—here it comes—an “NBA-style playoff” involving best-of-seven formats, etc. Of course, the charm of March Madness would be gone.
But the comparison does not hold for football. Schools in BCS conferences really need to win their league titles, and small conference teams had better run the table. The regular season remains extremely important because the margin for error is so thin. Under the arrangement we have now, a team from a BCS conference can have a perfect record (read: Auburn) and still not get a shot at the title. So for teams like Auburn in 2004—and virtually ALL small conference teams EVERY year—the entire regular season means nothing whatsoever.
Explain to me how a playoff diminishes THAT.