College Football - An Analysis of the BCS Conferences
A standard issue for debate amongst college football fans is the relative strength of conferences. It can generate alot of heat, but the argument is seldom accompanied by much in the way of actual facts.
For some fans, it's not enough for others to admit that their conferences has been perhaps the strongest over the last decade. No, their conference has to be the best every single year. And not only the best, it has to be acknowledged as clearly the best. This is particularly true for fans of one conference in particular. You know who you are.
So, here are some actual facts for the debate this season. My apologies for the small size.
This table displays:
1. The name of a conference. The 6 major conferences are included.
2. Their record against teams from the other 5 major conferences.
3. Their total points for and total points against in said games, as well as the average margin of victory (or defeat in those games).
4. The relative schedule difficulty for the conference in said games. Note: a negative score means their opponents have had better conference records than they themselves had, and a positive score means the opposite. For instance, if the average ACC team playing a team from major conference had a ACC record of 4-4, then their average opponent had a 4.55-3.45 record and their conference.
5. The conference's record in games as a favorite. Favorite is defined as their team having a conference record one game or more better than their opponent's conference record in their conference. For example, Ohio St. at (7-1) in the Big 10, vs. Washington at (2-7) in the PAC-10 counts as a favorite game for the Big 10.
6. The conference's record in games that are even match-ups. Even match-ups are defined as games in which the teams have identical conference records, or in which the teams have conference records that are within a half game of each other. For example, Virginia at (6-2) in the ACC vs. Connecticut at (5-2) in the Big East counts as an even game for both conferences.
7. The conference's record in games as an underdog. Underdog is defined as their team having a conference record one game or more worse than their opponent's conference record in their conference. For example, Ohio St. at (7-1) in the Big 10, vs. Washington at (2-7) in the PAC-10 counts as a underdog game for the PAC
10.
Notes:
1. Conference championship games are not counted in determining a team's conference record. This is because a conference championship game artificially strengthens a conference vs. a conference that does not have one (by giving a top team an additional conference loss, which is more significant than a top team getting an additional win, it would lower the expectation of the loser in out-of-conference play more than it would raise the winner's).
2. Although it may look like the the PAC-10 is ducking other conferences more than other conferences are doing, it has to be remembered that the PAC-10 plays an extra conference game than other conferences (two more than the Big East), in effect giving them another 10 games against BCS level opponents that other conferences don't play. Also, the Big East plays one less conference game than the other conferences (two less than the PAC 10), in effect giving them 8 less games against BCS level competition than other conferences.
3. This is obviously an attempt to compare conferences based on how they fared in games against each other. It is not perfect for two reasons. The first is obvious, teams do not always play at the same level, they have better games and worse games. This is an unavoidable problem for any way of analysis (polls or anything else) and can't be fixed by any method.
The second is that the ACC, Big 10, Big 12, and SEC have unbalanced schedules. That is, a team in those conferences does not play every other team in that conference. This somewhat balances itself out. It could be fixed by treating separate schedules as separate conferences, i.e. the group of teams Kansas played (also including Kansas) could be labeled Big 12a, and the group of teams Oklahoma played (also including Oklahoma) could be labeled Big 12b, and so on. However, this would be horribly complicated and it would be difficult to speak of a conference as a single entity.
Another possible solution would be to treat the two divisions of conferences as separate conferences, with the games between schools in different divisions not counting or counting as out of conference games. Either way, you would have to speak of the conference as 2 entities and not 1 (which is necessary for the purpose of this analysis). In addition, this approach would not work for the Big 10, which does not have divisions but does have an unbalanced schedule.
4. Notre Dame (and hence games against Notre Dame) are not included in this study, even though they are an independent BCS level team (albeit a poor one this season). This is because it would be unfair to do so when some conferences have played games against Notre Dame and others haven't.
It could also be misleading including Notre Dame and not other schools from mid-majors that are nevertheless of major conference quality, such as Hawaii and BYU (meaning that they could compete with major conference schools better than Baylor or Mississippi, not that they would have their same record if they played in a major conference). This study looks solely at games between schools from the 6 major (BCS) conferences.
The conferences ranked by the various means:
Winning Percentage:
1. PAC 10: .625
2. Big 10: .556
3. SEC: .500
4. ACC: .474
5. Big East: .467
6. Big 12: .455
Scoring Margin:
1. Big 10: 6.00
2. PAC 10: 5.25
3. SEC: 3.79
4. Big East: 1.07
5. Big 12: -3.00
6. ACC: -6.95
Schedule Difficulty:
1. PAC 10: -0.63
2. ACC: -0.55
3. SEC: -0.29
4. Big 12: -0.18
5. Big East: 0.1
6. Big 10: 2.22
Winning Percentage as Favorite:
1. Big East: 1.00
2. SEC: .800
3. Big 12: .750
4. PAC 10: .667
5. ACC: .600
6. Big 10: .571
Winning Percentage as Even:
1. Big 10: 1.00*
1. PAC 10: 1.00*
3. ACC: .667
4. SEC: .500
5. Big East: .286
6. Big 12: .000*
Winning Percentage as Underdog:
1. PAC 10: .500
2. Big 12: .333
3. ACC: .250
3. Big East: .250
5. SEC: .200
6. Big 10: .000*
* - donates that the conference only played one game in that category
An attempt to form a semi-objective synthesis:
Adding the ranks of winning percentage, scoring margin, and schedule difficulty gives:
1. PAC 10: 4
2. Big 10: 9
2. SEC: 9
4. ACC: 12
5. Big East: 14
6. Big 12: 15
A second attempt to form a semi-objective synthesis:
The formula is the same as above, however winning percentage is replaced by the average ranks of the winning percentage in the 3 categories (favorite, even, and underdog).
1. PAC 10: 5.17
2. SEC: 9.67
3. Big 10: 11.5
4. ACC: 11.83
5. Big East: 12.17
6. Big 12: 12.67
Criticism of the attempts to form a semi-objective synthesis:
First, it makes the classic analyst error of giving categories equal weight. However, there is no reason to conclude that the categories should be weighted equally. Secondly, the ranks are evenly spaced, but the actual data is not. For example, the difference between the schedule difficulties of first and second place is smaller than the difference between second and third place, but the difference between their ranks is not. Thirdly, the categories are not mutually exclusive, but the extent to which they are related is not possible to determine (although one might be able to make an approximation if you had access to all the data in a supercomputer).
The second attempt has an additional criticism. In two categories there are conferences that only played one game. But who is to say that the PAC-10's 1-0 record in even games is more impressive than the ACC's 4-2 record. An alternative may be to take games over or above .500, but this creates other problems.
My subjective rankings:
1. PAC-10: They had the best overall record, and it was against the toughest average match-ups of any conference. They also had the 2nd best average scoring margin. They are the obvious top pick for the best conference in games between major conference teams.
2. SEC: In some ways they seem merely a perfectly average conference. A 7-7 record, a perfect mirror - 4-1 as a favorite, 2-2 as evens, and 1-4 as underdogs. However, they did this with a positive scoring margin and a unfavorable average match-up, thus earning my second spot.
After these two, deciding between the conferences gets really difficult. There is simply no way to account for the Big 10's amazingly favorable match-ups, their strength of schedule number is the most unique number in the whole study. This makes evaluated the conference virtually impossible.
The average Big 10 team playing a non-conference game had a conference record of 4.89-3.11, whereas their opponents average conference record was 2.67-5.33. What this means is that the Big 10's top teams played more games against other major conferences than their bottom teams did, and they almost invariably played against the bottom half of other conferences. Of their 9 games against other major conference teams, their were favored (by respective conference records) in 7 of them, and only a underdog in one. Despite having an overall record of 5-4, and having the highest margin of victory, they had the worse record of any conference as a favorite at 4-3, and only played one game in the other two categories, which makes forming a conclusion about those categories difficult.
Compare this to the ACC which went 9-10 overall and had the worst scoring margin, but faced the second toughest match-ups and had arguably the best performance in even match-ups. Therefore, I'm wussing out and making a tie for third.
3. ACC
3. Big 10
5. Big East - Despite their 2-5 record in even games, their impressive 4-0 record as favorites, and their positive score margin gives them the nod to avoid the cellar.
6. Big 12 - My pick for last, the combination of the worst record and second worse scoring margin puts them in the basement.
The Bottom Line
In conclusion, the thing that stands out the most is how even the conferences are. There's decent arguments for having the 3rd-6th conferences in any order, and even the PAC-10 was only one game from falling back to the pack (horrible pun intended). The claims of homer fans of a certain conference (once again, you know who you are) that their conference is far and away the best are unfounded.
For some fans, it's not enough for others to admit that their conferences has been perhaps the strongest over the last decade. No, their conference has to be the best every single year. And not only the best, it has to be acknowledged as clearly the best. This is particularly true for fans of one conference in particular. You know who you are.
So, here are some actual facts for the debate this season. My apologies for the small size.
This table displays:
1. The name of a conference. The 6 major conferences are included.
2. Their record against teams from the other 5 major conferences.
3. Their total points for and total points against in said games, as well as the average margin of victory (or defeat in those games).
4. The relative schedule difficulty for the conference in said games. Note: a negative score means their opponents have had better conference records than they themselves had, and a positive score means the opposite. For instance, if the average ACC team playing a team from major conference had a ACC record of 4-4, then their average opponent had a 4.55-3.45 record and their conference.
5. The conference's record in games as a favorite. Favorite is defined as their team having a conference record one game or more better than their opponent's conference record in their conference. For example, Ohio St. at (7-1) in the Big 10, vs. Washington at (2-7) in the PAC-10 counts as a favorite game for the Big 10.
6. The conference's record in games that are even match-ups. Even match-ups are defined as games in which the teams have identical conference records, or in which the teams have conference records that are within a half game of each other. For example, Virginia at (6-2) in the ACC vs. Connecticut at (5-2) in the Big East counts as an even game for both conferences.
7. The conference's record in games as an underdog. Underdog is defined as their team having a conference record one game or more worse than their opponent's conference record in their conference. For example, Ohio St. at (7-1) in the Big 10, vs. Washington at (2-7) in the PAC-10 counts as a underdog game for the PAC
10.
Notes:
1. Conference championship games are not counted in determining a team's conference record. This is because a conference championship game artificially strengthens a conference vs. a conference that does not have one (by giving a top team an additional conference loss, which is more significant than a top team getting an additional win, it would lower the expectation of the loser in out-of-conference play more than it would raise the winner's).
2. Although it may look like the the PAC-10 is ducking other conferences more than other conferences are doing, it has to be remembered that the PAC-10 plays an extra conference game than other conferences (two more than the Big East), in effect giving them another 10 games against BCS level opponents that other conferences don't play. Also, the Big East plays one less conference game than the other conferences (two less than the PAC 10), in effect giving them 8 less games against BCS level competition than other conferences.
3. This is obviously an attempt to compare conferences based on how they fared in games against each other. It is not perfect for two reasons. The first is obvious, teams do not always play at the same level, they have better games and worse games. This is an unavoidable problem for any way of analysis (polls or anything else) and can't be fixed by any method.
The second is that the ACC, Big 10, Big 12, and SEC have unbalanced schedules. That is, a team in those conferences does not play every other team in that conference. This somewhat balances itself out. It could be fixed by treating separate schedules as separate conferences, i.e. the group of teams Kansas played (also including Kansas) could be labeled Big 12a, and the group of teams Oklahoma played (also including Oklahoma) could be labeled Big 12b, and so on. However, this would be horribly complicated and it would be difficult to speak of a conference as a single entity.
Another possible solution would be to treat the two divisions of conferences as separate conferences, with the games between schools in different divisions not counting or counting as out of conference games. Either way, you would have to speak of the conference as 2 entities and not 1 (which is necessary for the purpose of this analysis). In addition, this approach would not work for the Big 10, which does not have divisions but does have an unbalanced schedule.
4. Notre Dame (and hence games against Notre Dame) are not included in this study, even though they are an independent BCS level team (albeit a poor one this season). This is because it would be unfair to do so when some conferences have played games against Notre Dame and others haven't.
It could also be misleading including Notre Dame and not other schools from mid-majors that are nevertheless of major conference quality, such as Hawaii and BYU (meaning that they could compete with major conference schools better than Baylor or Mississippi, not that they would have their same record if they played in a major conference). This study looks solely at games between schools from the 6 major (BCS) conferences.
The conferences ranked by the various means:
Winning Percentage:
1. PAC 10: .625
2. Big 10: .556
3. SEC: .500
4. ACC: .474
5. Big East: .467
6. Big 12: .455
Scoring Margin:
1. Big 10: 6.00
2. PAC 10: 5.25
3. SEC: 3.79
4. Big East: 1.07
5. Big 12: -3.00
6. ACC: -6.95
Schedule Difficulty:
1. PAC 10: -0.63
2. ACC: -0.55
3. SEC: -0.29
4. Big 12: -0.18
5. Big East: 0.1
6. Big 10: 2.22
Winning Percentage as Favorite:
1. Big East: 1.00
2. SEC: .800
3. Big 12: .750
4. PAC 10: .667
5. ACC: .600
6. Big 10: .571
Winning Percentage as Even:
1. Big 10: 1.00*
1. PAC 10: 1.00*
3. ACC: .667
4. SEC: .500
5. Big East: .286
6. Big 12: .000*
Winning Percentage as Underdog:
1. PAC 10: .500
2. Big 12: .333
3. ACC: .250
3. Big East: .250
5. SEC: .200
6. Big 10: .000*
* - donates that the conference only played one game in that category
An attempt to form a semi-objective synthesis:
Adding the ranks of winning percentage, scoring margin, and schedule difficulty gives:
1. PAC 10: 4
2. Big 10: 9
2. SEC: 9
4. ACC: 12
5. Big East: 14
6. Big 12: 15
A second attempt to form a semi-objective synthesis:
The formula is the same as above, however winning percentage is replaced by the average ranks of the winning percentage in the 3 categories (favorite, even, and underdog).
1. PAC 10: 5.17
2. SEC: 9.67
3. Big 10: 11.5
4. ACC: 11.83
5. Big East: 12.17
6. Big 12: 12.67
Criticism of the attempts to form a semi-objective synthesis:
First, it makes the classic analyst error of giving categories equal weight. However, there is no reason to conclude that the categories should be weighted equally. Secondly, the ranks are evenly spaced, but the actual data is not. For example, the difference between the schedule difficulties of first and second place is smaller than the difference between second and third place, but the difference between their ranks is not. Thirdly, the categories are not mutually exclusive, but the extent to which they are related is not possible to determine (although one might be able to make an approximation if you had access to all the data in a supercomputer).
The second attempt has an additional criticism. In two categories there are conferences that only played one game. But who is to say that the PAC-10's 1-0 record in even games is more impressive than the ACC's 4-2 record. An alternative may be to take games over or above .500, but this creates other problems.
My subjective rankings:
1. PAC-10: They had the best overall record, and it was against the toughest average match-ups of any conference. They also had the 2nd best average scoring margin. They are the obvious top pick for the best conference in games between major conference teams.
2. SEC: In some ways they seem merely a perfectly average conference. A 7-7 record, a perfect mirror - 4-1 as a favorite, 2-2 as evens, and 1-4 as underdogs. However, they did this with a positive scoring margin and a unfavorable average match-up, thus earning my second spot.
After these two, deciding between the conferences gets really difficult. There is simply no way to account for the Big 10's amazingly favorable match-ups, their strength of schedule number is the most unique number in the whole study. This makes evaluated the conference virtually impossible.
The average Big 10 team playing a non-conference game had a conference record of 4.89-3.11, whereas their opponents average conference record was 2.67-5.33. What this means is that the Big 10's top teams played more games against other major conferences than their bottom teams did, and they almost invariably played against the bottom half of other conferences. Of their 9 games against other major conference teams, their were favored (by respective conference records) in 7 of them, and only a underdog in one. Despite having an overall record of 5-4, and having the highest margin of victory, they had the worse record of any conference as a favorite at 4-3, and only played one game in the other two categories, which makes forming a conclusion about those categories difficult.
Compare this to the ACC which went 9-10 overall and had the worst scoring margin, but faced the second toughest match-ups and had arguably the best performance in even match-ups. Therefore, I'm wussing out and making a tie for third.
3. ACC
3. Big 10
5. Big East - Despite their 2-5 record in even games, their impressive 4-0 record as favorites, and their positive score margin gives them the nod to avoid the cellar.
6. Big 12 - My pick for last, the combination of the worst record and second worse scoring margin puts them in the basement.
The Bottom Line
In conclusion, the thing that stands out the most is how even the conferences are. There's decent arguments for having the 3rd-6th conferences in any order, and even the PAC-10 was only one game from falling back to the pack (horrible pun intended). The claims of homer fans of a certain conference (once again, you know who you are) that their conference is far and away the best are unfounded.
posted by Freethinker at 11:21:00 PM
10 Comments:
Interesting... however, as a non-SEC fan, I recognize that the SEC is by FAR the best conference, especially in depth. And as a Big 10 fan, I realize that it is one of the worst conferences. The Pac-10 is vastly overrated, and Arizona State will prove that. The Pac-10 is still the same conference it was 10 years ago that can't play defense and can't beat anyone worthwhile outside of its own conference, with the exception being USC.
My own fact-less ranking:
1. SEC - Florida, LSU, Georgia, Kentucky, Arkansas... need I say more?
2. Big 12 - Nowhere as deep as the SEC, but the top guns of OU, Missouri, and Texas could play with the top SEC teams.
3. Pac 10 - USC and a healthy Oregon are solid, everyone else is really bad.
4. Big East - West Virginia is perenially overrated, and the rest of the conference stinks.
5. Big 10 - Will rebound, but this year they were awful. Ohio State is a big fish in a very small pond, and will get annihilated by LSU, as they did by Florida last year.
6. ACC - Name one good team from this conference? Va Tech is very overrated, having no offense and a mostly mediocre defense. If they play anyone with a pulse, they lose (for example, LSU beating them 48-7). Easily the worst conference.
I don't consider Kentucky and Arkansas to be good teams - sure they beat LSU - but Colorado beat Oklahoma and that doesn't make them a good team. The only non-conference team of note Kentucky played was a game at home against a bad Louisville team, a game they won by 6 points. Arkansas played possibly the most shameless non-conference schedule (Troy, North Texas, Chattanooga, Florida International, they did win all of those, so you can say Arkansas would be the best team in the Sun Belt) - and they have zero passing game.
You can't make statements about how good a conference is based on games they play against each other. Arkansas beats LSU - does that mean Arkansas is underrated or LSU is overrated? There's no way to tell really.
If the SEC were far and away the best conference you would expect that to show in their games against other BCS conferences, but it did not. You're right to question the defenses in the PAC-10, but the SEC defenses have not been impressive this season either.
Ok, I grant you that the SEC defenses are overrated. And Kentucky is probably overrated.
I don't think that one can draw conclusions from the FEW inter-BCS conference games. They are statistically insignificant, unless you take like a ten year span, which I think is hard to do with the ups and downs of conferences.
Bowl games are also a weak indicator, as one team from the Big Ten will take a bowl game seriously while its opponent from the SEC doesn't show up to play (Wisconsin vs. Arkansas last year). Same with the Gophers versus Texas Tech for 3 quarters last year; no way was Minnesota a 5-TD better team for the first 75% of that game.
Also, you have to consider HOW a team loses to its opponents. For example, LSU played fairly decently against Kentucky and very well against Arkansas (at least on offense), and were just beaten by better teams on that particular day (LSU is better overall, but Ark. and Kentucky gave them their best shot). Meanwhile, OU didn't give even close to their best shot against Colorado (or Iowa State), instead deciding to look ahead to the Texas game.
You're right about the few number of games, but I do think it might be enough to show there's no great differences between the BCS conferences. And bowl games are tough because of the 5-7 week lay-off.
It's very difficult though to tell the "how" as you put it. I mean, did Auburn and Alabama also play their best game against LSU (Auburn should have won, and Alabama easily could have), or is LSU perhaps simply not that good. Keep in mind, this is a an Auburn team that loss at home to South Florida, and an Alabama team that loss to Florida St. (both teams 4-4 in their conference) and even Louisiana-Monroe. I happen to think that LSU has a bad case of playing down to the competition, but who knows.
I think LSU is the Ohio State of the SEC... tons of talent, not really great at any one thing (though Ohio State at least plays good defense). LSU just throws athletes out there and hopes one of them makes a play or two. Flynn is an awful QB, but so is Boeckmann. I foresee a 42-17 LSU win. Ohio State doesn't have any weapons on offense that scare LSU, not after they've had to face Woodson and McFadden. And Ohio State's defense hasn't seen speed like LSU's offense since last year's championship game. And we all know how that turned out for Ohio State.
Well, so far the Big 12 is showing itself to be the best conference. Missouri destroyed an Arkansas team that ran all over LSU. Looks like Stoops was right, they should be in the championship.
Jinx!
Stoops better figure out a new motivational speech, that is 4 straight bowl games where the Sooners have shown up flatter than a pancake. They are incredibly more talented than West Va, but once again, their opponent wanted it more.
Funny, Pete Fiutak called the onside kick in his "Stream-of-Consciousness" on Collegefootballnews.com. Stoops is getting obvious in his coaching.
Post a Comment
<< Home