EBS Seedings and Matchups

Overview

Teams qualifying for the EBS tournament are seeded based on their relative positions in the rankings, with the exception of the four teams earning a first round bye who are always seeded 1-4. The remaining teams are seeded 5-12 according to their position in the rankings with no special benefit given to conference champions. First-round matchups are set based on the seeding following traditional tournament formats (i.e., 5 vs 12, 6 vs 11, 7 vs 10, and 8 vs 9). Exceptions are only made to avoid intra-conference matchups and repeats of regular season games. Later round matchups always follow traditional tournament formats without exceptions.

Methods

The matchups set in the EBS are driven by the following objectives:

Preserve the integrity of the bracket.
Avoid conference matchups.
Avoid regular season rematches.
Reward better teams with projected "easier" competition.

Details

Seeding

As described in the "Awarding of Byes" document, the four teams earning byes are seeded 1-4. The remaining seeds 5-12 are assigned according to the relative position of those teams in the final rankings. No preference is given to the method by which a team qualified for the EBS (i.e., whether or not the team was a conference champion).

Rankings

The Rankings used here refer to a mathematical construct designed to order the teams (as fairly as possible) based on their ability on the field. This system should include both the practical opinion of those closest to the sport as well as objective measures of the teams' performance.

The current BCS rating structure (enacted 2005) strikes this balance by combining two human polls and several computer measures. When examining the potential implementation of the EBS in past years, the current BCS formula was applied retroactively, as best as possible, in order to have consistency when comparing years. Any variations between the current rating structure and the one used in a given year are specified on the rating table for that year.

First-Round Matchups

For the most part, matchups are based on the traditional tournament format. In the first round, the #5 seed plays the #12 seed, #6 plays #11, #7 plays #10, and #8 plays #9. Exceptions to this are only made if the two teams slated to play one another are either both in the same conference or if the teams met during the regular season. Exceptions will only occur if changing matchups would not disadvantage too greatly another team in the bracket. From 1998-2006, potential first round conflicts occurred in only in five games.

Avoiding potential conference matchups or regular season rematches in the first round is considered more important than maintaining the integrity of the bracket for several reasons. First, the bowls traditionally seek to pit teams from differing parts of the country against one another. Having two teams from the same conference meet in a bowl goes against this tradition. Further, regional matchups, such as those resulting from intra-conference matchups, may not be appealing to a national audience and would hurt television ratings and interest in the postseason in general. Intra-conference matchups would also prohibit multiple teams from the same conference from advancing to later rounds of the tournament, and thus deny that conference a fair opportunity to demonstrate its worth compared to other conferences. Finally, rematches may provide an advantage to the team that lost the prior meeting, presumably the lower seeded team, and it would therefore limit the advantage earned by the higher seeded team in the EBS tournament.

When it may be necessary to change a matchup, only lower seeded teams may be switched. The higher seeds always maintain their position in the bracket. This is because the rating averages of the lower seeds are considered less precise than those of the higher seeded teams. When multiple conflicts arise, a hierarchy of undesirable games is used to determine switching. The priority list of games needing to be switched, in order, is as follows:

conference championship rematches
regular season conference rematches
conference teams who did not meet in the regular season
inter-conference rematches

When the team needing to be switched is the #9 seed, that team should be switched with the #10 seed, provided switching will not create another equal or higher-level conflict and the two teams have a rating average difference (as per the current BCS formula) no greater than 0.100. When the team needing to be switched is the #12 seed, that team should be switched with the #11 seed, provided switching will not create another equal or higher-level conflict and the two teams have a rating average difference no greater than 0.100.

When the team needing to be switched is seeded #10 or #11, that team may be switched with either the team immediately above or immediately below it in the seedings. It is preferable that the team be switched with the team whose rating average is nearest its own, provided switching will not create another equal or higher-level conflict, and the two teams have a rating average difference no greater than 0.100.

Two teams separated by two seeding spots may never trade places; however, as a last resort, a three-way switch (one team moving up two positions and then two teams each dropping one position) is permissible provided the team moving two positions is taking the place of a higher seeded team and the rating average difference between the two dropping teams and the teams they are replacing is no greater than 0.100.

Case Study of Potential First-Round Switches from the BCS Era

In 1999 Michigan was seeded #8 and Michigan St. was seeded #9. These two teams are both from the Big Ten Conference and met in the regular season. For this reason Michigan St. was switched with #10 seed Florida who was slated to play #7 Kansas St. This switch is permissible because it neither setup a conflict where conference teams would play one another nor does it lead to a regular season rematch in either game. Further, the switch is permissible because the difference between Michigan State's 0.699 rating average and Florida's 0.623 rating average was 0.076 (below the 0.100 threshold), meaning there was not a significant competitive disadvantage given to Kansas St. for having to play a higher seeded team.

In 2000, #8 Nebraska was slated to play #9 Kansas St. in the first round. Both these teams are in the Big 12 Conference and had met in the regular season. Kansas St. was switched with #10 Notre Dame who was slated to play #7 Florida. The difference between their rankings averages (KSU's 0.629 - ND's 0.553 was 0.076) was less than 0.100; however, Notre Dame had also played Nebraska during the regular season. Since Notre Dame and Nebraska are not in the same conference, the switch was permitted as intersectional rematches are preferable to conference rematches.

A potential conflict that could not be resolved occurred in 2001. #6 Florida was slated to play #11 LSU. LSU could not be switched with #10 Oklahoma because the #7 seed was Tennessee, another SEC conference opponent and a team LSU had played twice already that year. LSU could also not be switched with #12 BYU as the difference between their rating average (0.523 and 0.214, respectively, a 0.309 difference) was greater than 0.100 and would force Nebraska to play a significantly higher quality opponent. Likewise, the three-way switch could not be implemented. While #11 LSU could take #9 Texas' position without issue, and Texas could take #10 Oklahoma's spot without issue, Oklahoma could not move to LSU's spot since their rating average difference (0.651 and 0.523, respectively, difference of 0.128) was greater than 0.100, thus putting Florida at a disadvantage by having to play a significantly higher rated opponent. For these reasons, the matchups were left unchanged.

In 2002, another potential conflict could not be resolved. #8 Kansas St. was slated to play #9 Texas, two conference opponents who had met in the regular season. While there were no issues with switching Texas with #10 Notre Dame from a rating average perspective, the switch would have resulted in Texas having an intra-conference rematch with #7 Oklahoma. Because a team may not drop two ranking spots and because both Kansas St. and Oklahoma were regular season opponents of Texas, the matchups were left unchanged.

In 2003, #8 Tennessee was slated to play #9 Miami (FL) in the first round, but these two teams played an intersectional matchup during the regular season. For this reason, Miami (FL) was switched with #10 Kansas St., who was scheduled to play #7 Ohio St. Because neither teams were conference foes nor had they met in the regular season, and because the difference between the Miami (FL) and Kansas St. rating averages (0.675 and 0.636, respectively, difference of 0.039) was less than 0.100, the switch was permitted.

In 2007, #12 Florida would ordinarily play #5 Georgia; however, because these teams are from the same conference and met in the regular season, Florida and #11 Arizona St. were switched in the brackets. Their rating averages differed by only 0.007. Arizona St. now plays Georgia and Florida plays #6 Missouri.

Finally in 2009, #12 LSU would play #5 Florida according to traditional tournament matchups, but because these two teams are from the same conference and met in the regular season, LSU is switched with #11 Virginia Tech and will play #6 Boise St. instead. LSU and Virginia Tech differ by 0.031 rating points.

Quarterfinal & Semifinal Matchups

Quarterfinal matchups always follow traditional tournament format. The #1 seed plays the winner of the #8, and typically, #9 game; the #2 seed plays the winner of the #7/#10 game; the #3 seed plays the winner of the #6/#11 game; and the #4 seed plays the winner of the #5/#12 game. Matchups are not altered as these matchups, at the time of scheduling, would only be potential. It is advantageous for the bracket to be fixed once it has been announced and not to rearrange matchups based on the outcomes of first round games. Further, the higher seeds are thought to be more precise, and therefore, the fidelity to the bracket more important.

Semifinal matchups, likewise, follow traditional tournament format. The winner of the #1 vs #8/#9 game plays the winner of the #4 vs #5/#12 game. The winner of the #2 vs #7/#10 game plays the winner of the #3 vs #6/#11 game.