Conference Champions Only? A Playoff Case Study
DoubleB and I were/are engaged in a spirited debate under Brian's post regarding the home site concept being dead. Our debate is about whether or not to require entrants in the national title game to win their conference.
I believe the conference championship requirement is an important one for fairness and for preservation of the regular season. If you take a straight top four, you render many of the best regular season games (like Alabama-LSU last year, Michigan-Ohio State in 2006, USC-Notre Dame in 2005, etc) meaningless. Instead of being the biggest moments of the season, they are the least important. That, to me, is a crime against college football, where the regular season is more exciting than the playoffs of most sports.
DoubleB has made some good points against that idea, but he inadvertently introduced a piece of evidence that completely destroys the position: The 2008 college football season.
Here is the final BCS top ten from 2008:
1. Oklahoma (11-1)
2. Florida (11-1)
3. Texas (11-1)
4. Alabama (12-1)
5. USC (11-1)
6. Utah (12-0)
7. Texas Tech (11-1)
8. Penn State (11-1)
9. Boise State (12-0)
10. Ohio State (10-2) Terrell Who?
The final tallies of the AP and Harris polls had the same top four; the coaches poll ranked USC ahead of Alabama.
A four-team playoff constructed using BCS ranking criteria, taking the top four teams only, would give us a semi-final round featuring only Big 12 and SEC teams. It would probably look like this:
Fiesta Bowl:
#1 Oklahoma vs. #4 Alabama
Sugar Bowl:
#2 Florida vs. #3 Texas
This would be met with cries of injustice, bias, and corruption. And the first two critiques would be spot on. In this scenario USC and Utah are left out in the cold so that the "cool" conferences can get their second members. The problem is that the rankings here are just plain wrong. How do we know?
Oh, right.
2008 is a classic example of poll bias; pundits that know about as much as you and me watch football, think they know who looks good and who doesn't, and fill out polls that reflect their opinions. In 2008 everybody believed that the Big 12 and the SEC were the two best conferences. There seemed to be no question about it.
And everybody was wrong.
Now it may be that Florida was the best team in the country, but it's impossible to know for sure--Utah beat Alabama more convincingly than Florida did, and USC was absolutely unstoppable by the end of the season, as they were every year at the height of the Pete Carroll era. Unfortunately, we never saw the USC dynasty play a top SEC team during the mid-'00s. They did humiliate Auburn at home in 2003, but that Auburn team was a serious disappointment.
For all we know, USC was the best team in the country that year. Their only loss was early, on the road, to a talented Oregon State team; the next week they beat Oregon 44-10 in a game nobody noticed at the time, but looks a lot better now that we see that Chip Kelly was (as OC at the time) building Oregon into a powerhouse. This is the USC team that crushed Ohio State in Los Angeles 35-3; Texas needed every minute of the Fiesta Bowl to escape the same team. They defeated Penn State handily in the Rose Bowl. They were very good.
In the other direction, the Big 12 was already well on its way to becoming the defense-free league that nobody respected when Oklahoma State was begging for a Championship Game bid. It was a lot weaker than anybody wanted to believe, because they didn't have all the information.
And, of course, nobody believed Utah was good because they didn't even play in a "major" conference. No way they'd be able to handle the Big, Bad SEC.
Here's why it matters: There is no way to fairly rank teams based only on results, because there simply aren't enough results in a season where each team plays four non-conference games. There are biases that are present in the mind of every selector, every voter, every pundit.
Right now, for example, everybody believes that the SEC is far and away the top league; that may be right now, but it's not necessarily always true. And as much as they believe that, they have looked down on the Big Ten for decades. Even seasons when the B1G demonstrates its superiority on the field (1999, 2002) the story is buried because it doesn't fit in with current biases.
By requiring entrants to be conference champions, you help insure against those biases by preventing a love affair from a single conference from infecting the selection.
Would it work? Any time you test a theory like this, it's useful to apply it to past seasons to see how they would resolve. Let's apply it to the final 2008 standings and see what we get. Teams are selected based on ranking with teams that aren't conference champions disqualified:
1. Oklahoma (11-1)
2. Florida (11-1)
3. Texas (11-1)
4. Alabama (11-1)
5. USC (11-1)
6. Utah (12-0)
For a final seeding of:
1. Oklahoma
2. Florida
3. USC
4. Utah
That's much better. Fair. Just. Accurate. Compelling.
DoubleB adeptly provided a counterexample to the conference champion argument, that if LSU lost to Georgia in the SEC championship last season it's possible that neither of the best teams would be in the playoff. That is a legitimate criticism, but to fix the problem the SEC needs only to reform its championship structure to eliminate divisions and allow Bama and LSU to play each other. Alternatively, a compromise is available: Exchange the "Conference Champion" requirement for a "One team per conference" rule. That rule would preserve 2008 as I have adjusted it.
Verdict: In a four-team playoff, only conference champions should be admitted; or, at the least, only one team per conference.
Funny how the two financial imperatives -- larger and larger superconferences, and their consequent championship games -- enable, and also complicate, the operation of a playoff.
It is impossible to escape the inevitable conclusion that the conference commissioners' desires for a conference playoff, and the national sportswriters' guild's demand that there be a college football playoff, in any form, all detracts from the regular season (and home games in general).
So a hypothetical Michigan team en route to a national championship could play its last four games in Columbus, Indianapolis, Pasadena, and Tempe?
I think there are two concerns, however. One is that your n = 1. I think you'd need to look at a few years to give more weight to the argument. Two is that no matter what you do, hindsight will color your vision. So if we'd had the playoff system in place, the Utes never beat the Tide, and it's impossible to know how good the Tide were because they were left out of the playoff. I could see an argument down the road that yes, Utah won the championship, but only because they never had to play that awesome Alabama/Texas team. Not to mention the screaming if Utah had been put into the playoff and then lost the first game badly to the #1 seed.
You do have me moving more toward the concept of having only one team per conference, however, but primarily because of the potential for bias and the lack of significant inter-divisional play in order to give any significant statistical weight to won/loss records.
will NEVER work. Each conference will have its own idea of crowning a champion. Last year we were one bad call away from a losing-record UCLA from winning the Pac10 Championship. Also Georgia was a dropped pass away from beating LSU. So making it mandatory that the playoff teams be conference champions is useless in this age of Conference Championship Games. It will not be who is the best team in the confernece, it will be who is the luckiest team in the conference.
The only way to make it "fair" is to eliminate the Championship game and force everyone in the conference to play each other. Then if you have a head-to-head tie, the tie breaker will be used. No one could say they did not have a chance.
Haha, what??
What conference championship games did you watch??
LSU 42 Georgia 10
Oregon 49 UCLA 31
my point. Any number of lucky (or unlucky) plays would have resulted in a different outcome. Is that what you want? A UCLA team below .500 as a Conference Champion? It will happen.
What if Oregon got hit with suspension like they should have? Huh? Say they lose half of their starters. Then what?
I dont think so. I dont want to take the chance of the Conference Champion that is not the best team in the Conference.
that lucky breaks are exactly what happens in the Men's bball tourney all the time. That's what makes it so compelling to watch. Anything can happen.
There are ways to get deserving conference champions into the tournament while still allowing deserving non-champions into the tournament as well.
ALTON'S HYBRID SYSTEM: Take the final BCS ratings, but add .150 points to conference champion teams.
This way, the #1 team is still pretty much guaranteed a spot in the tournament whether or not they are conference champions. So in the example you posted, #5 USC and #6 Utah would have moved up to the third and fourth spots, and would make the tournament ahead of Texas and Alabama.
This year's rankings would have gone:
#1 LSU (1.150)
#2 Oklahoma State (1.083)
#3 Alabama (.942)
#4 Oregon (.940)
#5 Stanford (.848)
#6 Wisconsin (.787)
#7 Arkansas (.769)
#8 Boise State (.741)
So Oklahoma State moves up from #3 to #2, Oregon moves up from #5 to #4, and Wisconsin moves up from #10 to #6. I think this is more fair than champions only (which would have replaced Alabama with Wisconsin), and more fair than pure rankings (which would have replaced Oregon with Stanford).
Pure rankings: #1 LSU v #4 Stanford; #2 Alabama v #3 Oklahoma State
Champions only: #1 LSU v #4 Wisconsin; #2 Oklahoma State v #3 Oregon
Hybrid: #1 LSU v #4 Oregon; #2 Oklahoma State v #3 Alabama
I would take "Hybrid" over either of the other two.
The last 8 years, under the "Hybrid" system (I agree, it needs a better name):
2004: #1 USC v #4 Utah (up from #6), #2 Oklahoma v #3 Auburn
2005: #1 USC v #4 Ohio State, #2 Texas v #3 Penn State
2006: #1 Ohio State v #4 Louisville (up from #6), #2 Florida v #3 USC (up from #5)
2007: #1 Ohio State v #4 Oklahoma, #2 LSU v #3 Virginia Tech
2008: #1 Oklahoma v #4 Utah (up from #6), #2 Florida v #3 USC (up from #5)
2009: #1 Alabama v #4 Texas Christian, #2 Texas v #3 Cincinnati
2010: #1 Auburn v #4 Wisconsin (up from #5), #2 Oregon v #3 Texas Christian
2011: #1 LSU v #4 Oregon (up from #5), #2 Oklahoma State v #3 Alabama
Conference representation: SEC-8 teams, Big 12-6 teams, Pac 10/Pac 12-6 teams, Big Ten-5 teams, Mountain West-4 teams, Big East-2 teams, ACC-1 team.
I'm not crazy about the 2006 result under this system, but as Stephen pointed out in the original post, you can't call it unfair.
I don't think this idea would fly, but the best ideas often don't.
Perhaps BCS conference champs (excluding the Big East these days?) get the .150 bump, while non-BCS conference champs get a .100 bump?
The better idea, IMO, is to use strength of schedule. I'd rather give the .150 bump to a Boise team that managed to schedule four ranked teams in the OOC season and played none in conference compared to an ACC team winning their conference with three ranked wins and a MAC/FCS type OOC.
You'd end up with the SEC, B1G, Pac, BXII champions getting those points 99/100 years, but the last two spots could be up for grabs giving lower level conferences incentive to schedule tougher games.
I have no problem with bonus points for playing (and winning) particularly difficult games, but frankly you lose me when you use Boise State as an example of a team that should benefit. Boise State had the easiest overall schedule of any of the top 10 teams last year--it's just that their most difficult games were in the non-conference schedule rather than the conference schedule. I see no particular virtue in that, compared to teams that played 3 or 4 top-20 teams in conference.
You can substitute TCU, BYU, Hawaii, or one of the independants. I wasn't claiming Boise had some special claim on those bonus points.
This is a good plan, beautiful and full of win. Name needs work though.
better the chance to win a national title.
Makes perfect sense to me.
I've retyped my response to your response in the other thread:
In 2008, there were NINE teams that had 0 or 1 loss. The two 0-loss teams were Boise and Utah. There were 7 one-loss teams from major conferences: USC, Penn State, Florida, Bama, Texas, Texas Tech, and Oklahoma. In a 4-team playoff, multiple teams are getting screwed. That's just the reality of the situation. No matter the metric, some teams are taking it on the chin.
"This is a perfect example of poll bias"
It's not poll bias. Voters genuinely thought these were the best teams and it's tough to argue they weren't. Alabama was #1 until the SEC title game. Oklahoma was scoring 60 points a game at the end of the year, Florida won EVERY game from their loss to the SEC title game by 30+. Texas lost their game on the last play of the game. The computers LOVED the Big XII South. Had it been the computers with a 4-team playoff it would have included Oklahoma, Texas, Florida, and Texas Tech.
I do believe that in a 4-team playoff, there would have been some serious support for USC in that 4th spot in 2008. Voters weren't reevaluating their ballot at the 4 and 5 spots at that time. If they knew it counted, they would have made a more serious evaluation of Alabama and USC (or anybody else).
"people believed the Big 12 and the SEC were superior to other conferences, and they werewrong.
I don't think people believed that I all. I think they thought Florida, Oklahoma, and Texas were the best 3 teams in the country (all 3 polls and the computers agreed on this), USC and Bama were 4 and 5 in some order (the computers hated USC that year, the final pre-bowl Coaches poll had them actually tied), and then everyone else with some disagreement between the polls and computers. They voted for the teams, not the reputations.
The injustice is going to be having an NCAA tournament with #3 Stanford, #7 Virginia Tech, #10 Cincinnati, and #17 Michigan State because a couple of teams lose their conference title game.
A couple of quick responses:
You say that multiple teams are getting the the short end regardless; this is true, but it makes it all the more important to make sure that as many decisions as possible are made on the field; that is, if a team wins a competition against another team (winning a conference, for example) that team ought to get preference over the team on the short end. If you include fewer conferences you are introducing more subjectivity, and the injustice is far greater in magnitude.
"Poll bias" is just a term I'm using to describe the bias of voters, and unfortunately the fact is that a voter's belief about the quality of a team is inextricably linked to how strong they believe that team's conference is; this is why Boise State has not played for a national championship.
You argue that USC would probably have gotten some poll love for #4 in 2008; I agree, there is a good chance that they would have been able to sneak into that last spot. Good point.
Your supposed scenario is unrealistic; there is no existing scenario where that would have happened, and at any rate several compromises have been proposed (and supported) that would completely eliminate the chance of that happening. Your unlikely "nightmare scenario" is one of the reasons I am explicitly advocating compromise here.
Your arguments overall are contentious but fairly made. It's a useful way to stir up the issues involved.
because a couple of teams lose their conference title gameis pretty much the name of competitive sports across the board. I don't think if #17 MSU pounds the highest-rated team in the semifinal, that they will remain #17. Besides, #17 base on what? If they can beat a champion, then they are a champion. The Atlanta Braves were the second-best team in baseball in 1983 with a .299 team batting average, but they happened to be in the same division with the best team in baseball in a time before wild cards. Is this unfair? At least they could have gotten in by beating the LA Dodgers a few more times. What do you say to a team that never had a chance because Gary Danielson doesn't like (ie, isn't paid by) their conference? Get the competition out of the polls and on to the field. You want to be #1, then win your conference.
You can not have both, a playoff and a popularity contest. People want a playoff, but want to make sure the 4 best teams (opinion) have a chance at winning the championship. With confrence champions only, you turn tough divisional games into a playoff game. Because, if you don't win your division your not going to have a shot at making the 4 team playoff at the end of the season. Go 12-0 and lose your confrence championship game, which would now become part of the playoff, that sucks. There's no debate that can be made, everyone knows your team was better, but your team lost the wrong game on the field, not the popularity contest after the season.
In no other sport does it guarantee that the FOUR BEST teams are playing for a shot to win a championship. The Giants were not the best team in the NFL last year, not even one of the top 4. They probbly should not have even been in any playoff (opinion). They gave up more points than they scored, they didn't even have a winning record in thier division or confrence. How many people tuned in for the Super Bowl? On the other hand the Packers dominated the regulare season and won thier division only to lose to the Giants when it mattered most. The best team lost the wrong game and that is exactly what a playoff is, it's hardly ever the teams who people think are the best teams.
It can only be a four team popularity contest or a four team playoff it can not be both. Once the NCAA ditches the idea of popularity contests then they can start actually thinking playoff. Until then, its the same thing we have now, just add more fan bases into the mix who think they got screwed by the system.
Yes, I love the idea of a playoff because it determines who's the best team when it matters. Not who's the best team when there is nothing on the line.
That's why you need 8 teams, so that you get a good representation of top teams, and then let them prove who is the best when it matters.
If you are only going to have 4 teams, then there should only be one team from each conference. The teams who did not win their conference did not get it done when it mattered.
Though I don't think you need 8 teams. Six teams with #1 and #2 getting a bye would also do. But the 8-team idea has the virtue of allowing two non-BCS conferences to send their champs.
The key, as you note, is "performance when it counts." Using only results from on the field means that there is no advantage in terms of getting to the MNCG to scheduling cupcakes out of conference. OOC games that prepare for the conference would be the optimal competitive strategy, though money would still dictate some OOC matchups.
If using conference champions means sometimes getting the 7-6 UCLA-type teams into the playoffs, that's okay. If they win their conference, they have proven, by definition, that they deserve to be there.
I would love conference champions only.
Conference champions. I mean....what do conference non-champions have to even argue in their favor? They didn't win when it mattered? They lost at home in the biggest game of the year?
A lot of people probably assume I mean Alabama. I might be talking about Michigan 2006. That's the problem with college football memory.
You win or you don't win. It's simple. We aren't crowning "team that will win 51 or more out of 100 times," we're crowning National Champion. Put up or shut up. Bring it, don't sing it. Third cliche phase.
In 2006, "The Game" was in Columbus.
I agree there loosing at home in the biggest game was very disappointing.
An eight-team playoff would be a large enough field to include not only the conference champions, but also the highly ranked non-conference champions as well. If Iwere putting together a playoff today with the way college football is currently structured, it'd be the five major conference champions (ACC, Big Ten, Big XII, Pac 12, SEC) plus three at large teams in the field. The one caveat I'd add was that if a conference champion isn't in the top 12 of the rating system utilized, then it's replaced by an at large team. This would have happenened to ACC Champion Clemson last year using the BCS rankings.
For 2008, the ACC would not have been in the playoff because Virginia Tech (9-4) was ranked 19th in the BCS poll. That means four conference championship teams--Oklahoma (12-1), Florida (12-1), USC (11-1) and Penn State (11-1) would have been in the playoff with four at large teams. Those four programs would have been Texas (11-1), Alabama (12-1), Utah (12-0) and Texas Tech (11-1). At #9 in the polls, undefeated Boise State (12-0) would have been outside of the playoffs.
If these first round games were played at the stadiums of the higher rated teams, then here's the lineup of games:
#8 Penn State (11-1, B10 Champion) at #1 Oklahoma (12-1, Big XII Champion)
#5 USC (11-1, Pac 10 Champion) at #4 Alabama (12-1, SEC At Large)
#7 Texas Tech (11-1, Big XII At Large) at #2 Florida (12-1, SEC Champion)
# 6 Utah (12-0, MWC Champion) at #3 Texas (11-1, Big XII At Large)
If winning a conference championship is going to be used for post-season seeding with the top four conference champions getting the 1 thru 4 seeds, then the opening round playoff games would look like this:
#8 Penn State (11-1, B10 Champion) at #1 Oklahoma (12-1, Big XII Champion)
#5 Texas (11-1, Big XII At Large) at #4 Utah (12-0, MWC Champion)
#7 Texas Tech (11-1, Big XII At Large) at #2 Florida (12-1 SEC Champion)
#6 Alabama (12-1, SEC At Large) at #3 USC (11-1, Pac 10 Champion)
Instead of Alabama and Texas hosting first round games, Utah and USC would be doing it in this post season setup. This would be an award for winning a regular season conference championship and further ensure that the regular season results do translate into the post-season.
While Boise State would be left on the outside looking in, using eight teams instead of four better insures that most of the national championship capable teams are in the playoffs. Ultimate, what most fans want to see is the best teams that emerge from the regular season decide the championship out on the field with less reliance on committees, polls and computers to make the decisions on who participates.
Also with eight teams, the major bowls would have plenty of teams in the existing inventory to stage games against name teams. The programs listed 9 thru 16 in the BCS polls were Boise State, Ohio State, TCU, Cincinnati, Oklahoma State, Georgia Tech, Georgia and BYU. Oregon and Virginia Tech were 17 and 19 in the BCS poll.
Eventually, we'll see Division 1-A college football adapt some sort of eight-team playoff. It may come out thru realignment and restructuring as the conference championship games become de facto quater finals. The other alternative is something like the one I outlined above that makes sure the conferences get some skin in the game by giving the major champions autobids, but still ensure the better teams also participate with the minimum offering of three at-large bids.
This is interesting:
http://espn.go.com/college-football/story/_/id/7945482/big-12-sec-champ…
This means everything has been decided. The debate is over. These kinds of secondary decisions are not made until the primary decisions are made.
This allows us to deduce that it is conference champions only (with an exception for higher-ranked independents). This game (Sugar Bowl?) is where the future 2011 Alabamas will go, just like the future 2006 Michgans will be in the Rose Bowl.
The polls are proven to be incorrect nearly every week, yet there are those that insist on using them to determine who is deserving of selection into this or that. Obviously, people aren't satisfied by this.
The BCS should get out of the ranking business as much as it can. It should limit the teams it has to rank by selecting only from conference champions (and probably limit it's conferences too, but thats an argument for another day).
Conference 'champions 'are the only solution. If conferences want to let 'false' champions like UCLA advance (had it hypthetically won), it's on them. If they want to advance the highest ranked team - they can do that too.Nothing should stops a conferencce from choosing it's 'champion' based on polls, votes, wins or whatever arbitrary random criteria they want. If they screw it up, at least the playoff system won't catch all the heat. Their job gets much easier.
If the two best teams are in one conference they should play each other. If there is some doubt after that and you need a rematch - it should happen in the conference championship game. It's not the BCS's job to settle these things.
If every conference had a championship game, then a conference-winners-only playoff is just a division-winners only playoff that's one round longer with the first round always coming against another team from your conference.
The only playoff system in college football that would not detract from the regular season is a conference champion only playoff, and that is not debateable IMO. That includes the current faux one game playoff we have now--Alabama getting to play LSU again after losing to them at home in the regular season, rendered the regular season matchup between the two teams entirely meaningless. And that would have been the same in 2006 with us and OSU--which is why I was totally fine then with us not being put in the title game. A win over OSU would have made their victory over us meaningless as well. And that is bad for the make or break, every loss is crushing, regular season we have now, which makes this game unique.
nice
Comments