Tournament of Champions - 2006 Edition

We had so much fun with the 2007 Tournament of Champions, we thought we'd go back in time to look at how much fun the playoff could have been in prior years. If you haven't read it yet, make sure to look at the rationale behind the Tournament of Champions.

For 2006, we have a very interesting field. We'll use the BCS standings as our ranking/seeding system, and allow any independent that finishes #6 or higher an invitation.

Conference Champions:

Seeding	Team	Conference
1	Ohio St.	Big 10
2	Florida	SEC
3	USC	PAC 10
4	Louisville	Big East
5	Boise St.	WAC
6	Oklahoma	Big 12
7	Wake Forest	ACC
8	BYU	Mountain West
9	Houston	Conference USA
10	Central Michigan	MAC
11	Troy	Sun Belt

Independents: None qualify. Notre Dame finished #11 in the BCS standings.

The championship is played in the Fiesta Bowl, and in preserving the bowl system as much as possible, there could be terrific games for the the other top-tier bowls:

Rose: Michigan (11-1) vs. California (9-3)
Orange: Auburn (10-2) vs. Notre Dame (10-2)
Sugar: LSU (10-2) vs. Wisconsin (11-1)

The seeding for the 2006 Tournament of Champions produces the following bracket:

Notable in 2006:

• As in 2007, Ohio St. secures the #1 seed.
• The WAC champion (Boise St.) seeds better than the ACC champion (Wake Forest). This format does not take for granted that BCS conferences should have all the top seeds.
• Second round USC vs. Oklahoma would be a fun game. So would third-round Florida vs. USC.
  

And the National Champion is...

Boise State had an amazing season and proved in its real-life victory over Big 12 champion Oklahoma that it could hang with the big guys. As flat as Ohio St. was in its real-life bowl game, we'll take Boise St. to the finals with victories over Louisville and the Buckeyes. As strong as the Gators looked in real life, we don't think anyone beats them in this fictional post-season. Florida cruises past Wake, has a reasonable victory over the Trojans, and ends the Broncos' Cinderella season.

Tournament of Champions - 2007 Edition

The Tournament of Champions is a playoff alternative to the current BCS system that secures the participation of all 11 conference champions in Division 1-A college football to establish a divisional champion. In addition to crowning a genuine national champion, the Tournament aims to preserve the bowl tradition, ensuring (by not taking second-place conference finishers) that top-tier bowls keep top-notch teams.

The following demonstrates how this proposal would work following the 2007 regular season. We'll use the BCS standings as our ranking/seeding system, and allow any independent that finishes #6 or higher an invitation.

Conference Champions:

Seeding	Team	Conference
1	Ohio St.	Big 10
2	LSU	SEC
3	Virginia Tech	ACC
4	Oklahoma	Big 12
5	USC	PAC 10
6	West Virginia	Big East
7	Hawaii	WAC
8	BYU	Mountain West
9	Central Florida	Conference USA
10	Central Michigan	MAC
11	Florida Atlantic	Sun Belt

Independents: None qualify.

With the championship held in the Sugar Bowl, the other top-tier bowls have terrific games to be had among the #2 teams in each conference, with selections possibly going like this:

Rose: Arizona State vs. Illinois
Orange: Georgia vs. Kansas (or Boston College?)
Fiesta: Missouri vs. Florida

Each of these top bowl games would be worth the ticket. As for the Tournament of Champions, the brackets would be as follows:

Notable:

• #4 Oklahoma vs. #5 USC would be a killer early round game.
  
• With Ohio State looking flat toward the end of the season, an upset to BYU is entirely possible.
• Hawaii's undefeated season probably ends at #2 LSU.
  
• West Virginia vs. Virginia Tech would be a great regional matchup.
  

And the National Champion is...

We think Oklahoma, squeaking past USC in a game-for-the-ages, comes away with this championship beating BYU handily, then LSU in a close one. No matter who emerged from the Ohio St./Oklahoma/USC bracket, there would be no dispute that they belonged in the championship game.

For that matter, Hawaii winning over LSU and Virginia Tech would prove it belonged as well. Are we so sure that wouldn't happen, if the championship were allowed to be settled on the field rather than in polls and computers?

The Tournament of Champions

In previous posts, I've argued that the BCS as currently structured does not solve the problem of determining a division 1-A college football champion, primarily because it is exclusive to only a subset of division 1-A teams and conferences, and because it affords post-season eligibility to too few teams upon completion of the regular season. A playoff alternative of 8 to 16 teams certainly preserves the value of the regular season, with discussion typically focusing on one of two broad options:

• An 8-team tournament: Six champions from among the BCS conference schools, plus two at-large selections
  
  
• A 16-team tournament: Eleven champions from among all Division 1-A conferences, plus five at-large selections
  

Each has its pros and cons. The 8-team field is nice in that there's only two extra games for participating teams, and the smaller number allows for only conference champions and outstanding independents. Fewer teams also allow for top-level bowl games to enlist top-notch competition among the teams that do not make the tournament. Still, the two at-large selections may (and likely will) pass over deserving non-BCS conference teams in favor of one- or two-loss BCS teams.

A 16-team field ensures that all division 1-A conferences have a chance to participate. There is something very stabilizing about the rule "win your conference and you are in" as it takes the eligibility control away from pollsters and computers and puts it in the hands, feet, bodies, and brains of the teams themselves. For any team that would complain about not being included, there is a simple answer: win your conference next time. Five at-large selections ensures that outstanding independents can participate and affords a second chance to teams that may have been hit with injuries for a crucial game during the regular season. That said, the 16-team format does add a third extra game for all participating teams and the five at-large spots will surely take away any chance for top-tier bowls to invite top-notch competition. It's one thing for the Rose Bowl to be asked to accept the second-place Big 10 and PAC 10 teams; if first- and second-placed teams in those conferences happen to be invited to the tournament, does the Rose Bowl (and other top-tier bowls) settle for third?

Bowl games are a wonderful, unique tradition in this sport, and should (and could) be preserved even with a post-season tournament. In particular, the traditional January bowls - the Rose, Fiesta, Orange, and Sugar - should continue to enjoy top-tier competition. Going down the bowl chain, it probably doesn't matter so much. If the Alamo played a #5 Big 12 team rather than a #4, it's probably no worse off. But the top-tier bowls are certainly diminished if too many teams are selected for a national championship tournament, and that is undesirable. Certain other bowls which invite a non-BCS conference champion, such as the Liberty Bowl (it takes the Conference USA champion), would also be forced to adjust. Hopefully the #2 team in C-USA would provide just as good a game.

Still, even with their cons, each format featuring either an 8- or 16-team field is better than the current BCS format with the potential to crown a genuine Division 1-A national champion. Another playoff alternative could mitigate the problems for top-tier bowls and is worth consideration.

Call it the NCAA Division 1-A Tournament of Champions. Simply put, if you win your conference you're invited to play other champions for the National Championship. If you don't win your conference, you're not invited. The number of participants matches the number of conferences in Division 1-A, so the field presently is eleven. The teams are seeded either by committee or RPI-like-formula into a bracket like the following:

The top five champions are rewarded with a first-round bye, and as the tournament winner is likely to come from this group, the concern that participants will be subjected to too many extra games prior to the championship is mitigated - most likely the winner will have played only two. Even so, wouldn't lesser-seeded teams gladly play the extra game for a real chance at a national championship? This format succeeds in producing a genuine national championship among all division 1-A conferences.

Allow better-seeded teams a home field advantage in the first two rounds (maybe three) and there is tremendous incentive not just to win one's conference, but to do so in a manner as to secure the better post-season seed. This will certainly reward conference champions who elected to play tougher non-conference schedules. In fact, this format encourages teams to play reasonable out-of-conference opposition as scheduling non-conference cupcakes offers no reward. There is a huge difference between getting the 5th and 6th seeds here (for example) and stronger non-conference play would be a deciding factor.

For that matter, teams are neither rewarded nor penalized for the existence of a conference championship game. Missouri was hurt by a loss in the Big 12 championship. LSU was substantially helped by playing the extra game for their conference title. Ohio State was helped by not having to play one; had USC the opportunity to play (and win) a conference championship, they might be playing in the BCS title game. In the Tournament of Champions system, each conference gets to choose their representative however they wish without consequences for other conferences.

An additional strength to this format, assuming one wishes to preserve the existing bowl system as much as possible, is that it keeps all the second place conference finishers available for top-tier bowls. This is significant, for it ensures that there is excellent competition to be had even with the conference champions unavailable. Assuming the Sugar Bowl as the site for the National Championship, this year would see the likes of Georgia, Missouri, Kansas, Boston College, Illinois, and Arizona State playing as top competition in the Rose, Orange, and Fiesta Bowls. That is not much of a drop-off, if any at all, from what currently exists with this year's BCS selections. To further preserve the bowls as much as possible:

• Allow the tournament championship to be rotated as the BCS does now among the top-tier bowls.
  
  
• Keep most existing bowls and their conference tie-ins (without league champions, but including #2 on).
  
  
• Ensure the tournament games through semi-finals are played in the mid-December weeks so as not to interfere with a Bowl week following Christmas.
  
  
• Play the other top-tier bowls on January 1st to maintain tradition and play the National Championship on January 2nd.
  

Some would argue that a two-loss Georgia has more right to be eligible to play for a National Championship than say a five-loss Florida Atlantic. While it is true that 99 times out of 100, Georgia beats Florida Atlantic on a neutral field, limiting the field to conference champions makes the selection system entirely objective. A big part of the BCS problem is the arbitrary subjectivity of the selection system. Again, here the answer is simple: just win your conference. Teams that play in more difficult conferences but finish second are still rewarded with placement in top-tier bowls, and that is significant. The second-place team in the Sun Belt likely won't get invited to play in the Orange Bowl.

Besides, the argument that non-BCS teams have no chance to advance in a tournament just doesn't hold true universally. Could the 2004 Utah team have advanced to the finals? Absolutely. How about the 2006 Boise State Broncos? You bet they could have advanced. The beauty of selecting conference champions is that nothing is assumed - everything is decided on the field. The 2007 Hawaii Warriors could get blown out in the first round - or they could very well surprise many in a tournament like this. Yes, for the time being, the tournament champion would likely come from among the current BCS conferences, but this system ensures that isn't taken for granted.

If this Tournament of Champions has a weakness, it is that an eleven-team field affords no opportunity for independents to participate. That is an unacceptable deal-breaker. Even the current BCS includes Notre Dame, and with continued improvement, who's to say we won't see a one-loss Navy team in the near future? I admit I don't have an easy answer to this issue. While it simplifies the system dramatically to force these teams into conferences, that isn't likely to happen. A more workable solution is to allow a 12th team into the field under a very specific set of conditions. In the case of a twelve-team field, the 5th-seeded team loses its bye and hosts the 12th-seeded team in the first round. If no independent qualifies, the field remains at eleven.

One possibility: If an independent finishes higher in the ranking system than the highest-rated second-place conference team, then the independent is invited and seeded accordingly among a twelve-team field. If two-loss Georgia finishes second in the SEC but has a higher ranking than Notre Dame, the Irish aren't invited. If however Notre Dame (or any other independent) finishes higher than every other second-place conference team, they have proven they belong among the conference champions and are invited.

That might be a little too harsh a standard for independents. Even so, independents cannot be allowed special privilege just because they aren't competing in a conference. The selection must also be objective. A second possibility: any independent that finishes in the top #6 of the selection rankings is invited to the tournament. Top #10 is too many - this is a Tournament of Champions after all. It is likely that one second-place conference finisher will appear in the top #6, but perhaps as many as three or four in a top #10. So we'll pick a top #6 threshold for an independent to prove they belong among other conference champions.

An eleven-team field of conference champions in its objectivity provides an elegant path to a genuine division 1-A National Championship. As tournament formats go, it allows participation from all division 1-A conferences while the existing bowl system loses little. If a solution for the participation of independents were ironed out, a true division 1-A champion could be had.

Championship Eligibility: The Regular Season Matters

If ever a division 1-A college football season proved that the BCS diminishes the value of the regular season, it's this one.

Supporters of the BCS frequently argue that a post-season playoff for a national championship is inferior to the BCS formula of polls and computers. The argument goes something like this: in division 1-A football, The Regular Season Matters. Each game is a do-or-die situation for each team - so much so that the entire regular season is the playoff. The BCS, they argue, is superior to a playoff because it preserves the importance of the regular season. This makes division 1-A football unique among all sports, collegiate and professional. It makes our sport special.

The reality is that, with respect to national championship eligibility, the BCS structure belittles the significance of the regular season for a number of teams. Examples going back to the BCS's inception in 1998 abound. Consider the following division 1-A teams who completed their Regular Season undefeated but with no eligibility to compete for their divisional championship:

• 1998: Tulane, 11-0, Conference USA champions

• 2004: Auburn, 12-0, SEC champions

• 2004: Utah, 12-0, Mountain West champions

• 2004: Boise State, 11-0, WAC champions

• 2006: Boise State, 12-0, WAC champions

• 2007: Hawaii, 12-0, WAC champions

BCS supporters argue that competition in conferences like the WAC and C-USA is weaker than in the BCS conferences, so it is acceptable to exclude these teams from the BCS title game. Indeed, considering that the BCS is designed to produce a champion from 66 BCS conference teams (including Notre Dame), it is acceptable in this system to exclude them. This serves to underscore the primary flaw with the BCS - it produces a BCS champion, not a division 1-A champion among 119 teams and 11 conferences. If it is acceptable to have conferences like the WAC and C-USA competing in the division, then there is no excuse to exclude their undefeated champions from divisional championship eligibility. In denying their eligibility, the BCS diminishes the value of their regular seasons.

Of course, the regular season for the 2004 Auburn team, undefeated champions of the BCS Southeastern Conference didn't matter either. The BCS is designed to devalue the regular seasons of even BCS conference teams if it has to.

There are many other examples that illustrate how the BCS disregards the regular season for several teams.

• At the end of the 2000 season, one-loss Florida State was eligible for the national championship, despite being beaten in head-to-head regular season competition by Miami, who also finished with only one loss. Miami was not eligible for the national championship. For that matter, neither was 10-1 Washington who gave Miami their only loss.
  
  
• After 2001, one-loss Oregon was denied eligibility. So was the Big 12 champion Colorado. Selected instead was Nebraska who had previously lost to Colorado.
  
  
• After 2003, one-loss USC was denied eligibility, despite a #1 ranking in the AP poll and a PAC 10 championship. Big 12 runner up Oklahoma was selected instead, fresh off their crushing loss to Kansas State in the Big 12 championship game.
  
  
• After 2006, denying Michigan a chance to replay Ohio State for the national championship was reasonable, as it had lost the Big 10 to the Buckeyes. Hold the same true for one-loss Wisconsin. Less reasonable was denying eligibility to one-loss Louisville and undefeated Boise State.
  
  
• After the 2007 regular season, LSU and Ohio State were deemed eligible to play for the national championship, even though two-loss LSU was beat in its last regular season game. Strong finishes by ACC champion Viriginia Tech and Big 12 champion Oklahoma didn't matter. Neither did an undefeated season for Hawaii.
  
  

The problem with all this of course isn't so much the teams that are selected - there is always a solid case for their eligibility. Even in the 2003-04 debacle, the case could be made that Oklahoma's one loss shouldn't count worse for them than USC's - and USC after all didn't have to play the extra championship game that Oklahoma did. No, the problem is with the teams that are left out, for there is always a valid case for their eligibility as well. The BCS simply makes too few teams eligible for championship contention.

Furthermore, to apply the argument the regular season matters to avoid discussion of a playoff alternative is the strangest of red herrings. The undeniable truth is that without a good regular season, you don't win your conference. Without a good regular season, you don't qualify for the post-season. The regular season matters more if there is a post-season playoff than it currently does for the BCS.

There is one respectable point in the BCS-supporters' side of this argument. If too many teams are deemed eligible at season's end to compete for the national championship it dilutes the importance of the regular season. While that is fair, it is equally fair to say that deeming too few teams at season's end as eligible also tarnishes the value of the regular season. There is a sweet spot to be had here to keep the Regular Season on its appropriate pedestal - the right number of teams to select for divisional championship eligibility.

A useful metric for a post-season system is the ratio of the number of teams eligible for a national championship to the total number of teams competing in the regular season. Call it the "Championship Eligibility" ratio. For example, the NBA has 30 teams competing in its regular season, with 16 selected for post-season championship eligibility. That is a whopping 53.3% Championship Eligibility ratio. It would be fair to argue that with its post-season format, the NBA dilutes the importance of its regular season. At the other end of the spectrum is the currently structured BCS, featuring 66 teams competing in its regular season, with two selected from that group for championship eligibility. That is a paltry 3.0% ratio and reduces the value of its regular season just as much by not allowing enough teams. Comparing the ratios across different collegiate and professional leagues provides some insight for determining that sweet spot:

#RS = The number of teams competing in the regular season
#PS = The number of teams selected for post-season championship eligibility
% Eligible = The percent of teams eligible for post-season championship

League	#RS	#PS	% Eligible
NBA	30	16	53.3%
NHL	30	16	53.3%
NFL	32	12	37.5%
NCAA Div. 1 Men's Ice Hockey	59	16	27.1%
Major League Baseball	30	8	26.7%
NCAA Div. 1 Men's Soccer	202	48	23.8%
NCAA Div. 1 Softball	277	64	23.1%
NCAA Div. 1 Men's Baseball	293	64	21.8%
NCAA Div. 1 Field Hockey	78	16	20.5%
NCAA Div. 1 Women's Volleyball	324	64	19.8%
NCAA Div. 1 Men's Basketball	336	65	19.4%
NCAA Div. 1 Women's Basketball	335	64	19.1%
NCAA Div. 1-AA Football	111	16	14.4%
BCS	66	2	3.0%

The major professional leagues are all above 25% in this metric - that may be too high a ratio to keep the Regular Season as meaningful as BCS supporters would like. It seems 15% could be a reasonable sweet spot, with the 1-AA division in football comparing very favorably at 14.4%. Indeed, a 15% ratio is far lower than what is currently used in most NCAA sports; some may argue for a ratio as high as 20% for comparability to other Division 1 sports.

Here are the Championship Eligibility ratios for some of the more commonly suggested post-season scenarios for division 1-A football:

Div. 1-A Football Post-Season Scenario	#RS	#PS	% Eligible
Tournament: 16 seeded teams among Div. 1A	119	16	13.4%
Tournament: 12 seeded teams among Div. 1A	119	12	10.1%
Tournament: 11 seeded teams among Div. 1A	119	11	9.2%
Tournament: 8 seeded teams among Div. 1A	119	8	6.7%
BCS Championship with 4 eligible BCS teams	66	4	6.1%

A tournament with even as many as 16 teams providing a 13.4% eligibility ratio still comes in under the currently used post-season system for division 1-AA, which itself maintains a far lower ratio than the major professional leagues and other NCAA sports. While preferable to the current BCS format, the "BCS+1" format that establishes essentially a four-team playoff might still be a little low at only 6.1% eligibility.

Interestingly, and particularly relevant to the BCS supporters' argument: the ratios for the "BCS+1" 4-team format and an 8-team tournament among all 119 div. 1-A teams are almost the same at between 6 and 7%. To claim that the former preserves the value of the regular season while the latter doesn't is a false argument as they are statistically identical in this regard.

The more relevant argument to be had between a BCS+1 tournament and an 8-team division 1-A tournament is whether or not college football should have a champion only among BCS conference teams, or a champion among all 119 division 1-A teams. If you believe the championship belongs exclusively among the 66 BCS teams, then the BCS+1 playoff scenario is an excellent format and preserves the value of the regular season better than the current BCS.

If however you believe that national championship eligibility in division 1-A college football should incorporate all 119 teams, then feel comfortable with a post-season tournament among 8 to 16 teams. Championship Eligibility ratios between 6.7% and 13.4% certainly maintain the importance of The Regular Season.

The BCS: Solving the Wrong Problem

Since 1998, the month of December has seen one particular activity as fashionable as holiday shopping among college football fans: intense debate over the Bowl Championship Series. Like hearing the constant repetitions of tiresome carol arrangements played over department store intercoms, the ongoing BCS debate drudges along as part of a holiday season backdrop that must be tolerated, however unsatisfying it may be.

The debate typically centers around whether or not BCS "works" - whether in a given year it picks the appropriate two teams to play for the national championship. Each and every year since its inception, BCS team selections have been worthy of castigation, and teams with reasonable arguments for inclusion have been left out. This has fueled the heated, often bitter discussion among coaches, players, fans, and analysts.

Arguments over whether the BCS works however are fundamentally misplaced. There is nothing inherently wrong with the BCS. The real source of the systematic frustration surrounding the college football post-season is that the BCS is a solution to the wrong problem.

The reason for the creation of the BCS was described well by the then-BCS Chairman Kevin Weilberg responding in 2005 to a congressional hearing invitation for his testimony:

"The current structure is designed to match the No. 1 and 2 ranked teams, identified through a ranking system, in a bowl game. It is an extension of the bowl system and a method to determine a national champion through the bowls," Weiberg said. "It has paired teams in bowl games that would not have been possible under the bowl arrangements existing before its creation." [ref]

Weinberg's comments underscore the primary BCS function: pair two teams to compete for a declared national championship in a preserved bowl system. Significantly, the "national championship" part is less relevant than the "bowl preservation" part. The BCS is no more than a yearly adjustment of existing bowl arrangements to afford an arbitrary pairing of two teams.

For its stated purpose, the BCS does work. The heads of the six major conferences (plus Notre Dame) and their bowl executive partners created this arrangement to select (their best? their most deserving?) two teams to play in a bowl exhibition, with the winner crowned their BCS champion. The system addresses a very different problem than that of determining a Division 1-A champion among 11 conferences and 119 teams, and the difference isn't subtle.

The criticism heaped on the BCS year after year is almost unfair. Its purpose is not, nor has it ever been, to determine a Division 1-A champion among 11 conferences and 119 teams. Its goal is and has always been to match the No. 1 and 2 ranked teams among the six BCS conferences, "identified through a ranking system, in a bowl game." It doesn't really matter what the selection rules are, or how arbitrary they may appear in any given year. "Use human polls a and g this year, weighted at k% and throw in the average of computer polls x, y, and z... " As long as the six BCS conferences and their bowl executive partners agree to the rules, the BCS does its job and two teams play for its championship.

The appropriate problem to solve is the one for which the BCS isn't designed. Unless the NCAA is willing to restructure the 1-A division to include only the six BCS conferences, the correct issue to address is this: How to best determine a national champion among 11 conferences and 119 division 1-A teams?

It is no wonder that there has been controversy every year with BCS selections. Following a regular season, to deem only two teams as worthy of consideration for a national title out of 119 possibilities is ridiculous. Likewise, to limit the pool of worthy candidates to only those teams from six conferences is to say to the teams from the other five that they are not really competing in division 1-A.

Consider that with good regular season play, any division 1-AA football team may qualify for a post-season playoff and thus have a legitimate chance to win their national championship.
All teams in all major professional sports leagues are eligible for their respective league championships. For that matter, teams from any division 1-A conference in sports other than football are eligible with good regular season play to compete for their respective national championships.

Division 1-A football is lamentably unique among amateur and professional sports alike in that nearly half its teams are not even eligible to compete for their national title. Realistically, only teams from the six BCS conferences (and Notre Dame) can qualify. Others may play, indeed have played, in one of the prestigious BCS bowls, but the BCS national championship game will always be reserved for two teams from the ACC, Big East, Big 10, Big 12, PAC 10, and SEC conferences. Past (and present) BCS selections have proven that even an undefeated mid-major team will be passed over for a multiple-loss BCS conference participant. If division 1-A football is going to include the MAC, Conference USA, Mountain West, WAC, and Sun Belt conferences, teams from those conferences should be eligible for their divisional championship.

Successful problem solving begins with correctly identifying the problem. It is time for the NCAA to allow a broader field of worthy candidates the eligibility to participate in its division 1-A national championship process. The BCS must go. It simply solves the wrong problem.