To have a playoff, or stick with the BCS?
It is quite fashionable these days to regard the BCS as, at best, the simple manufacturing of a contest to award some team a mythical championship based primarily on the public perception(s) of those teams in the mix. At worst, the BCS represents a cash grab by the "power conferences" (ACC, Big East, Big Ten, Big XII, Pac 10, SEC) who have the tradition (and fan base) to warrant inclusion into BCS bowl games which they might not otherwise "deserve." A good example, though not in a BCS conference per se, is Notre Dame - which regularly plays in bowl games it might not truly "deserve" to be in simply for the fact that whichever bowl grabs them can be assured of filling the seats in the stadium. While this discussion recognizes that there are millions of dollars at issue here (and, more importantly, no doubt actually driving the decision of whether to go to a playoff or not), the working assumption for our purposes here is that whatever system is in place, people will make plenty of money. That is to say, the money issue is really more about who gets it and not how much of it is available. This piece doesn't care who gets the money. Some care will be taken to avoid the money issue in the posts below, though - as you'll see - there are times when the issue will be mentioned.
A brief history of the BCS might be appropriate here, but for the sake of brevity, it will not be outlined fully. For our purposes here keep in mind the following assumption: The BCS was created as a means to remove - or at least to put a check on - voter bias as well as to give the fans a game matching the teams perceived to be the top 2 in the nation a chance to settle it on the field. On some level, this is believed to take some of the "mythical" out of the "Mythical National Champion" label, but in practice has revealed itself to be as arbitrary as the former system of simply taking a poll. In other words, it is a rare season where only 2 teams identify themselves as a clear cut above the rest - 2002 being a true example of only 2 teams having an "honest claim" (and even then, many people think 2 loss USC was probably the best team). In a more common year, as many as 4 or 5 teams can argue their case as to why they "deserve" to be playing for all the marbles. While I maintain the BCS does precisely what it's designed to do - differentiate between otherwise like teams - I will not belabor the point in the following.
Instead, I will outline three separate ideas for a playoff so that we can see what Playoff proponents are really clamoring for. The first analysis in each post below will be based on the following principal: "Fairness" Proponents of the playoff system, or "anti-BCS-ers," often toss around the idea of fairness. This concept is two fold. 1 - fairness of competition. The power conferences have a stranglehold on the BCS games, and some people believe that the mid-major conferences should be considered more. Using 2008 as an example, Utah played in a conference, the MWC, which compares favorably to several so called "power" conferences and thus, Utah's running the table in the MWC should not be discounted. 2 - The unavoidable financial consideration. As above, I won't get in to these factors too much, but will mention that the "fairness" argument contemplates mid majors not just getting a piece of the BCS pie, but a piece of the BCS pie that goes along with a championship. That said, to be principled, we can't selectively decide the MWC gets a bite at the pie, but the Sun Belt does not. No. To rest on the principal of fairness, we must include ALL conference champions. Thus, as outlined below, the first analysis will contemplate what a playoff would have looked like using ONLY conference champions. As will, I think, be demonstrated, inviting conference champions, and conference champions alone, does little to generate "better" match-ups during the playoffs than the current bowl structure already generates. Such a formulation removes concepts of margin of victory and strength of schedule from consideration entirely. While some feel this is appropriate (and for reason), I believe it will be demonstrated below does not truly give college fans what they "want" in the long term.
Second on each post will be a 6+2 system. While proponents of the playoff system seem convinced it is the better way to determine a champion, few seem to be able to articulate exactly who should be invited. That said, one solution I've heard several times considers taking the champion of each power conference and then the remaining two at large teams available and regardless of conference affiliation. It is likewise unclear under what formula those two at large teams would be selected from. For our purposes here, we will keep the BCS formulation as a "final ranking" of sorts for said determination. The important point here, I think, is recognition that we are now moving away from principled reason for change, and instead simply masking the arbitrary nature of the BCS in a bigger pool of contestants. Likewise, as will be demonstrated below, in most years little is done to accommodate the mid-majors and their "fairness" concerns (both in terms of pure football skill and financial). While a different selection process could be conceived, if you hear someone argue 6+2 with the BCS rankings determining the at large bids, you should be aware you're not really basing your playoffs on a different principle and instead simply asking for "fairness" over some other metric - perceived skill on the field comparative to the rest of the nation. While the BCS limits this selection process to 2, the 6+2 system rests on the same flaws and gives us 8.
Finally, moving completely away from the principal of taking conference champions, we shall consider a playoff of the top 8 BCS teams in each season. Some proponents of this idea would take 4, others as many as 16. I settled on 8 because it is between those tallies and I believe will adequately demonstrate - illustratively - the benefits and problems of either a 4 or 16 team field. As mentioned, the "fairness" issue here is completely removed from conference affiliation, and instead speaks of fairness on a level of who is perceived as the top 8 in the nation. As a process, it suffers the exact same problems as the current BCS (ie being dependent in part on human bias, and computer programming issues), though it invites more teams to the party. As will be seen below, such a system does tend to generate more "great games" than does the current bowl system (at least at a glance), but also amounts to little more than the same money grab already present while resting on the same flaws as the current system. In short, while potentially more exciting over a number of games, there is little principle involved save for the assumption that humans and computers maybe can't select the top 2, but can figure out the top 8 adequately enough.
There will be some commentary in each post regarding the information, and I think it'll become clear (if it's not already) that there is little desire for a principled playoff (Conference champions) and instead people simply want to have better post season games to watch. Don't get me wrong, there is nothing wrong with that. The problem I have is people going beyond the assertion by suggesting that it's more "fair" or more "principled" or any of that. A 6+2 and an 8 team playoff is as arbitrary as the current BCS, masking the particular failings under a larger field, while at the same time offering the potential for a team that probably doesn't truly deserve consideration for the title, a shot at that title. That is to suggest that while a playoff does represent a system of determination of a champion, there are no benefits in that system which are remarkably different than the current system. While we would accept a playoff champion as the national champion, the truth is, the selection settles little and is just as "mythical" in reality. A playoff makes little effort to reward the best team, choosing the best team on a particular day against a particular opponent. It's as random as anything else. That is easily illustrated with the following example: Suppose #1 USC plays #4 Texas and wins 17-15 in an epic battle. #3 Utah plays #2 Alabama and wins 28-10. Can't Texas argue it would have beat Utah, and only lost as a consequence of the seeding? Of course they could. A playoff settles nothing. It's merely a recognized system of determination. It's legitimate. But, its not the only way to determine a champion and it is not fail safe.
Methodology
In setteling on what numbers to use I had to make some choices. It is recognized that there are other ways in which to evaluate the information and therefore the conclusions will be tainted by these choices. That said, we should agree that we need some manner to evaluate the data and so here are the reasons the information considered was choosen, and for what ends those means decided upon.
Conference champions - In the event Co-Champions from a conference, the team selected for our purposes will be by considering the team from that conference who went to the BCS bowl (ie Purdue in 2000 from the Big Ten, despite being Co-Champs with Michigan and Northwestern). If there was a Co-Championship in a conference which does not receive an auto-bid to a BCS game, the team selected was selected based on the following principle, in order: Overall Record, Higher BCS rank, stronger schedule, final AP rank. As before, it is recognized that this may not be the case in "reality" all the time, but for our purposes here, in the interest of full disclosure, was the manner of selection.
At Large Selections - At large selections were made by taking the highest BCS rating remaining in the pool of teams once conference champions had been removed (in the 6+2 system) or simply taking the top 8, as the BCS does not generate (or has not generated) ties.
Strength of Schedule - All SOS numbers are dervied from Sagarin's formulation. I choose this for two reasons. First, the datat was readily available and accessible. Second, while there are other formulations, the Sagarin SOS ratings consider a multitude of factors which I feel are germane for consideration. It is admitted that there are other means to determine SOS. To the extent, however, that any one process is chosen, the data is internally consistent by use of the same formula throughout each year. That said, it must also be noted, the SOS rankings below are post bowl rankings, which is perhaps inappropriate with respect to selection for our hypothetical playoffs. It is further noted, that in the year 2000, and only for this year, the SOS rankings are pre-bowl and I could not find post bowl rankings.
Final Sagarin Ratings - Like the selection of the AP poll, the Sagarin ratings were used for the limited purpose of illustrating what the computers "thought" of a team. There are several computer polls which could have been utilized for this purpose, some perhaps better and some perhaps worse. Again, the use of the Sagarin rating was in part owing to ease of finding the information (as it came on the same page as the SOS data). It is recognized that the poll suffers it's own failings - most notably demonstrated in 2002 where Sagarin listed USC as number 1, Miami of Florida 2, and Ohio State 3, despite the fact that Ohio State was 13-0 (USC was 11-2) and had beaten Miami on the field. That said, use of the poll is - again - not intended to illustrate any "truth of the matter" but to instead capture on some level the computer poll results.
Brackets - The number of conferences to be considered has changed since 1998. As such, the Conference Champions Alone formulation is inconsistent with respect to size - but consistent with regard to the number of conferences at issue in a particular season. The Sun Belt conference did not exist until 2001, for example. Likewise, the MWC did not come in to being until 1999 and the "Big West" conference ended play in 2000. First round byes were given to the teams with the highest BCS ranks to manufacture brackets when necessary. When constructing brackets, care was taken to give seeds 1 and 2 "easier" roads to the title (ie the NCAA tournament pairing 1 v. 16, 2 v. 15 and so on.) Thus, for example, in a 10 game bracket, 9 plays 10 with the winner playing 1, while 7 plays 8 with the winner playing 2. 4 plays 6 winner to face winner of 1 v. 10/9 game, 3 plays 5 with winner playing winner of 2 v. 7/8 game.
Why not a Plus One? - I have not considered the Plus One format because I think it patently absurd. I have no clue why we would be snowed in to thinking that after a 12 game season, we don't have enough information to pick Title Game contestants, but after bowl season, suddenly we have enough information to select the truly "worthy" contestants. The very idea of a Plus One isn't itself an end, but a means to making people (read, college Presidents and ADs) comfortable with a structure that sort of resembles a playoff. I could be wrong in that conclusion, but I can see no other purpose for it.
With that said, on to the data, and some commentary which I hope develops some of the themes outlined above:
** I have decided to make each year it's own separate thread. First, readers could be "turned off" of reading much of the information if presented in a single thread due to the sheer girth of information to be presented. Second, it occurs to me that each year (and this methodology) may merit discussion and critique and it seems to make more sense to dedicate that discussion to it's individual sub-topic rather than a single thread mish-mash.
Link to Methodology
Link to 1998 Playoff Hypotheticals
Link to 1999 Playoff Hypotheticals
Link to 2000 Playoff Hypotheticals
Link to 2001 Playoff Hypotheticals
Link to 2002 Playoff Hypotheticals
Link to 2003 Playoff Hypotheticals
Link to 2004 Playoff Hypotheticals
Link to 2005 Playoff Hypotheticals
Link to 2006 Playoff Hypotheticals
Link to 2007 Playoff Hypotheticals
Link to 2008 Playoff Hypotheticals
Link to Conclusions
It is quite fashionable these days to regard the BCS as, at best, the simple manufacturing of a contest to award some team a mythical championship based primarily on the public perception(s) of those teams in the mix. At worst, the BCS represents a cash grab by the "power conferences" (ACC, Big East, Big Ten, Big XII, Pac 10, SEC) who have the tradition (and fan base) to warrant inclusion into BCS bowl games which they might not otherwise "deserve." A good example, though not in a BCS conference per se, is Notre Dame - which regularly plays in bowl games it might not truly "deserve" to be in simply for the fact that whichever bowl grabs them can be assured of filling the seats in the stadium. While this discussion recognizes that there are millions of dollars at issue here (and, more importantly, no doubt actually driving the decision of whether to go to a playoff or not), the working assumption for our purposes here is that whatever system is in place, people will make plenty of money. That is to say, the money issue is really more about who gets it and not how much of it is available. This piece doesn't care who gets the money. Some care will be taken to avoid the money issue in the posts below, though - as you'll see - there are times when the issue will be mentioned.
A brief history of the BCS might be appropriate here, but for the sake of brevity, it will not be outlined fully. For our purposes here keep in mind the following assumption: The BCS was created as a means to remove - or at least to put a check on - voter bias as well as to give the fans a game matching the teams perceived to be the top 2 in the nation a chance to settle it on the field. On some level, this is believed to take some of the "mythical" out of the "Mythical National Champion" label, but in practice has revealed itself to be as arbitrary as the former system of simply taking a poll. In other words, it is a rare season where only 2 teams identify themselves as a clear cut above the rest - 2002 being a true example of only 2 teams having an "honest claim" (and even then, many people think 2 loss USC was probably the best team). In a more common year, as many as 4 or 5 teams can argue their case as to why they "deserve" to be playing for all the marbles. While I maintain the BCS does precisely what it's designed to do - differentiate between otherwise like teams - I will not belabor the point in the following.
Instead, I will outline three separate ideas for a playoff so that we can see what Playoff proponents are really clamoring for. The first analysis in each post below will be based on the following principal: "Fairness" Proponents of the playoff system, or "anti-BCS-ers," often toss around the idea of fairness. This concept is two fold. 1 - fairness of competition. The power conferences have a stranglehold on the BCS games, and some people believe that the mid-major conferences should be considered more. Using 2008 as an example, Utah played in a conference, the MWC, which compares favorably to several so called "power" conferences and thus, Utah's running the table in the MWC should not be discounted. 2 - The unavoidable financial consideration. As above, I won't get in to these factors too much, but will mention that the "fairness" argument contemplates mid majors not just getting a piece of the BCS pie, but a piece of the BCS pie that goes along with a championship. That said, to be principled, we can't selectively decide the MWC gets a bite at the pie, but the Sun Belt does not. No. To rest on the principal of fairness, we must include ALL conference champions. Thus, as outlined below, the first analysis will contemplate what a playoff would have looked like using ONLY conference champions. As will, I think, be demonstrated, inviting conference champions, and conference champions alone, does little to generate "better" match-ups during the playoffs than the current bowl structure already generates. Such a formulation removes concepts of margin of victory and strength of schedule from consideration entirely. While some feel this is appropriate (and for reason), I believe it will be demonstrated below does not truly give college fans what they "want" in the long term.
Second on each post will be a 6+2 system. While proponents of the playoff system seem convinced it is the better way to determine a champion, few seem to be able to articulate exactly who should be invited. That said, one solution I've heard several times considers taking the champion of each power conference and then the remaining two at large teams available and regardless of conference affiliation. It is likewise unclear under what formula those two at large teams would be selected from. For our purposes here, we will keep the BCS formulation as a "final ranking" of sorts for said determination. The important point here, I think, is recognition that we are now moving away from principled reason for change, and instead simply masking the arbitrary nature of the BCS in a bigger pool of contestants. Likewise, as will be demonstrated below, in most years little is done to accommodate the mid-majors and their "fairness" concerns (both in terms of pure football skill and financial). While a different selection process could be conceived, if you hear someone argue 6+2 with the BCS rankings determining the at large bids, you should be aware you're not really basing your playoffs on a different principle and instead simply asking for "fairness" over some other metric - perceived skill on the field comparative to the rest of the nation. While the BCS limits this selection process to 2, the 6+2 system rests on the same flaws and gives us 8.
Finally, moving completely away from the principal of taking conference champions, we shall consider a playoff of the top 8 BCS teams in each season. Some proponents of this idea would take 4, others as many as 16. I settled on 8 because it is between those tallies and I believe will adequately demonstrate - illustratively - the benefits and problems of either a 4 or 16 team field. As mentioned, the "fairness" issue here is completely removed from conference affiliation, and instead speaks of fairness on a level of who is perceived as the top 8 in the nation. As a process, it suffers the exact same problems as the current BCS (ie being dependent in part on human bias, and computer programming issues), though it invites more teams to the party. As will be seen below, such a system does tend to generate more "great games" than does the current bowl system (at least at a glance), but also amounts to little more than the same money grab already present while resting on the same flaws as the current system. In short, while potentially more exciting over a number of games, there is little principle involved save for the assumption that humans and computers maybe can't select the top 2, but can figure out the top 8 adequately enough.
There will be some commentary in each post regarding the information, and I think it'll become clear (if it's not already) that there is little desire for a principled playoff (Conference champions) and instead people simply want to have better post season games to watch. Don't get me wrong, there is nothing wrong with that. The problem I have is people going beyond the assertion by suggesting that it's more "fair" or more "principled" or any of that. A 6+2 and an 8 team playoff is as arbitrary as the current BCS, masking the particular failings under a larger field, while at the same time offering the potential for a team that probably doesn't truly deserve consideration for the title, a shot at that title. That is to suggest that while a playoff does represent a system of determination of a champion, there are no benefits in that system which are remarkably different than the current system. While we would accept a playoff champion as the national champion, the truth is, the selection settles little and is just as "mythical" in reality. A playoff makes little effort to reward the best team, choosing the best team on a particular day against a particular opponent. It's as random as anything else. That is easily illustrated with the following example: Suppose #1 USC plays #4 Texas and wins 17-15 in an epic battle. #3 Utah plays #2 Alabama and wins 28-10. Can't Texas argue it would have beat Utah, and only lost as a consequence of the seeding? Of course they could. A playoff settles nothing. It's merely a recognized system of determination. It's legitimate. But, its not the only way to determine a champion and it is not fail safe.
Methodology
In setteling on what numbers to use I had to make some choices. It is recognized that there are other ways in which to evaluate the information and therefore the conclusions will be tainted by these choices. That said, we should agree that we need some manner to evaluate the data and so here are the reasons the information considered was choosen, and for what ends those means decided upon.
Conference champions - In the event Co-Champions from a conference, the team selected for our purposes will be by considering the team from that conference who went to the BCS bowl (ie Purdue in 2000 from the Big Ten, despite being Co-Champs with Michigan and Northwestern). If there was a Co-Championship in a conference which does not receive an auto-bid to a BCS game, the team selected was selected based on the following principle, in order: Overall Record, Higher BCS rank, stronger schedule, final AP rank. As before, it is recognized that this may not be the case in "reality" all the time, but for our purposes here, in the interest of full disclosure, was the manner of selection.
At Large Selections - At large selections were made by taking the highest BCS rating remaining in the pool of teams once conference champions had been removed (in the 6+2 system) or simply taking the top 8, as the BCS does not generate (or has not generated) ties.
Strength of Schedule - All SOS numbers are dervied from Sagarin's formulation. I choose this for two reasons. First, the datat was readily available and accessible. Second, while there are other formulations, the Sagarin SOS ratings consider a multitude of factors which I feel are germane for consideration. It is admitted that there are other means to determine SOS. To the extent, however, that any one process is chosen, the data is internally consistent by use of the same formula throughout each year. That said, it must also be noted, the SOS rankings below are post bowl rankings, which is perhaps inappropriate with respect to selection for our hypothetical playoffs. It is further noted, that in the year 2000, and only for this year, the SOS rankings are pre-bowl and I could not find post bowl rankings.
Team Records - It must be noted, a team's record is post bowl. The reason for selecting post bowl record was merely for the ease of doing so, as it was contained on the same page(s) as the Sagarin data and I don't have unlimited time for this project. Pre-bowl records are probably more appropriate, but for our purposes here, I think, not a critical consideration.
Final AP - All AP rankings listed below are from the final AP poll for that season. It is recognized that the AP poll suffers several failings and is only useful for limited purposes. The intention herein is not to use the rating as an authority on a team's actual ability, but instead simply a guide as to the general "feeling" of how a team was regarded nationally. Thus, while reasonable minds may disagree with Ohio State ranking 4 in 2003, begind #3 Oklahoma, the data is not intended to establish OU was acutally better, but that they were perceived as such by one group of humans.
Final Sagarin Ratings - Like the selection of the AP poll, the Sagarin ratings were used for the limited purpose of illustrating what the computers "thought" of a team. There are several computer polls which could have been utilized for this purpose, some perhaps better and some perhaps worse. Again, the use of the Sagarin rating was in part owing to ease of finding the information (as it came on the same page as the SOS data). It is recognized that the poll suffers it's own failings - most notably demonstrated in 2002 where Sagarin listed USC as number 1, Miami of Florida 2, and Ohio State 3, despite the fact that Ohio State was 13-0 (USC was 11-2) and had beaten Miami on the field. That said, use of the poll is - again - not intended to illustrate any "truth of the matter" but to instead capture on some level the computer poll results.
Brackets - The number of conferences to be considered has changed since 1998. As such, the Conference Champions Alone formulation is inconsistent with respect to size - but consistent with regard to the number of conferences at issue in a particular season. The Sun Belt conference did not exist until 2001, for example. Likewise, the MWC did not come in to being until 1999 and the "Big West" conference ended play in 2000. First round byes were given to the teams with the highest BCS ranks to manufacture brackets when necessary. When constructing brackets, care was taken to give seeds 1 and 2 "easier" roads to the title (ie the NCAA tournament pairing 1 v. 16, 2 v. 15 and so on.) Thus, for example, in a 10 game bracket, 9 plays 10 with the winner playing 1, while 7 plays 8 with the winner playing 2. 4 plays 6 winner to face winner of 1 v. 10/9 game, 3 plays 5 with winner playing winner of 2 v. 7/8 game.
Seeding - The same principle in determining conference champions was used when determining seeding:Higher BCS rank, Overall Record, stronger schedule, final AP rank.
Errors - I have been careful to limit any errors in transcribing the information, but leave room for the occasional oversight. If you find an error, please report it to me and I will attempt to insert the correct data. Errors, however, should not be seen as having the hidden purpose of advancing my personal belief that a Playoff settles nothing.
Why not a Plus One? - I have not considered the Plus One format because I think it patently absurd. I have no clue why we would be snowed in to thinking that after a 12 game season, we don't have enough information to pick Title Game contestants, but after bowl season, suddenly we have enough information to select the truly "worthy" contestants. The very idea of a Plus One isn't itself an end, but a means to making people (read, college Presidents and ADs) comfortable with a structure that sort of resembles a playoff. I could be wrong in that conclusion, but I can see no other purpose for it.
With that said, on to the data, and some commentary which I hope develops some of the themes outlined above:
** I have decided to make each year it's own separate thread. First, readers could be "turned off" of reading much of the information if presented in a single thread due to the sheer girth of information to be presented. Second, it occurs to me that each year (and this methodology) may merit discussion and critique and it seems to make more sense to dedicate that discussion to it's individual sub-topic rather than a single thread mish-mash.
Link to Methodology
Link to 1998 Playoff Hypotheticals
Link to 1999 Playoff Hypotheticals
Link to 2000 Playoff Hypotheticals
Link to 2001 Playoff Hypotheticals
Link to 2002 Playoff Hypotheticals
Link to 2003 Playoff Hypotheticals
Link to 2004 Playoff Hypotheticals
Link to 2005 Playoff Hypotheticals
Link to 2006 Playoff Hypotheticals
Link to 2007 Playoff Hypotheticals
Link to 2008 Playoff Hypotheticals
Link to Conclusions
Last edited by a moderator:
