Analyzing the Sagarin Rankings

[xusandy]

Almost every team in the top conferences plays a bunch of games against cupcake teams in their pre-conference schedule, and (predictably) slaughters them. More interesting (and more important when dance cards are issued) is how teams from multi-bid conferences do against quality opponents. I studied the Sagarin rankings for a while today, and saw immediately that the Big East has a losing record (4-5) against the top 25 in the Sagarins, and also a losing record (10-11) against the top 50. But then I went to look at the other power conferences, and discovered that EVERY conference has a losing record against the top 25 and also against the top 50. Well, OK, the top 50 teams have lots of wins among them, and all of those wins count as a loss against a top 50 team for whomever they beat. So let's just look at how the top conferences have done against top 25 or top 50 opponents from other conferences.

Against the top 25 and top 50 teams in the Sagarins, and through games of 12/13, the records of the top 11 conferences in the Sagarins, listed below in order of their winning percentages against top 50 opposition:
1. Big East vs. top 25 = 4 wins and 5 losses, vs. top 50 = 10-11 (2.1 games per conference member against top 50 opposition)
2. Big 12 vs. top 25 = 4-6, vs. top 50 = 8-9 (1.7 games per member)
3. ACC vs. top 25 = 5-12, vs. top 50 = 14-17 (2.07 games per member)
4. Pac 12 vs. top 25 = 2-11, vs. top 50 = 8-15 (1.92 games per member)
5. Big 10 vs. top 25 = 4-21, vs. top 50 = 14-28 (3.0 games per member)
6. West Coast vs. top 25 = 2-3, vs. top 50 = 4-8 (1.2 games per member)
7. SEC vs. top 25 = 0-15 vs. top 50 = 9-25 (2.42 games per member)
8. Mo-Valley vs. top 25 = 1-6, vs. top 50 = 3-9 (1.2 games per member)
9. AAC vs. top 25 = 3-6, vs. top 50 = 5-21 (2.36 games per member)
10. Mountain West vs. top 25 = 0-5, vs. top 50 = 1-11 (1.09 games per member)
11. A-10 vs. top 25 = 1-9, vs. top 50 = 1-16 (1.06 games per member)

All other conferences combined have a grand total of 1 win against top 25 teams (IPFW over Indiana!), and just 2 more wins against the top 50 (1 each for Valpo and Texas-Arlington.) So we're clearly seeing a lot of pre-conference games this year pitting top teams against other top teams -- at least a lot more than happened even 2-3 years ago. Assuming that results to date are meaningful (and we're getting close to the start of the conference seasons), then:

(1) The Big East, Big 12, and ACC have the best winning percentages against top teams, with a gap back to the Pac 12, Big 10, and West Coast -- although Gonzaga skews the West Coast data significantly, since the Zags have 3 of that league's 4 wins against top 50 opponents.
(2) The Big 10 has collectively scheduled more games against top opponents per league member than anyone else. (I wonder why, but good for them!)
(3) No SEC team has yet won a game against a top 25 team, despite 15 such games (no kidding! huh?, wow!)
(4) As a league, the A-10 seems barely relevant.

[Bill Marsh]

Almost every team in the top conferences plays a bunch of games against cupcake teams in their pre-conference schedule, and (predictably) slaughters them. More interesting (and more important when dance cards are issued) is how teams from multi-bid conferences do against quality opponents. I studied the Sagarin rankings for a while today, and saw immediately that the Big East has a losing record (4-5) against the top 25 in the Sagarins, and also a losing record (10-11) against the top 50. But then I went to look at the other power conferences, and discovered that EVERY conference has a losing record against the top 25 and also against the top 50. Well, OK, the top 50 teams have lots of wins among them, and all of those wins count as a loss against a top 50 team for whomever they beat. So let's just look at how the top conferences have done against top 25 or top 50 opponents from other conferences.

Against the top 25 and top 50 teams in the Sagarins, and through games of 12/13, the records of the top 11 conferences in the Sagarins, listed below in order of their winning percentages against top 50 opposition:
1. Big East vs. top 25 = 4 wins and 5 losses, vs. top 50 = 10-11 (2.1 games per conference member against top 50 opposition)
2. Big 12 vs. top 25 = 4-6, vs. top 50 = 8-9 (1.7 games per member)
3. ACC vs. top 25 = 5-12, vs. top 50 = 14-17 (2.07 games per member)
4. Pac 12 vs. top 25 = 2-11, vs. top 50 = 8-15 (1.92 games per member)
5. Big 10 vs. top 25 = 4-21, vs. top 50 = 14-28 (3.0 games per member)
6. West Coast vs. top 25 = 2-3, vs. top 50 = 4-8 (1.2 games per member)
7. SEC vs. top 25 = 0-15 vs. top 50 = 9-25 (2.42 games per member)
8. Mo-Valley vs. top 25 = 1-6, vs. top 50 = 3-9 (1.2 games per member)
9. AAC vs. top 25 = 3-6, vs. top 50 = 5-21 (2.36 games per member)
10. Mountain West vs. top 25 = 0-5, vs. top 50 = 1-11 (1.09 games per member)
11. A-10 vs. top 25 = 1-9, vs. top 50 = 1-16 (1.06 games per member)

All other conferences combined have a grand total of 1 win against top 25 teams (IPFW over Indiana!), and just 2 more wins against the top 50 (1 each for Valpo and Texas-Arlington.) So we're clearly seeing a lot of pre-conference games this year pitting top teams against other top teams -- at least a lot more than happened even 2-3 years ago. Assuming that results to date are meaningful (and we're getting close to the start of the conference seasons), then:

(1) The Big East, Big 12, and ACC have the best winning percentages against top teams, with a gap back to the Pac 12, Big 10, and West Coast -- although Gonzaga skews the West Coast data significantly, since the Zags have 3 of that league's 4 wins against top 50 opponents.
(2) The Big 10 has collectively scheduled more games against top opponents per league member than anyone else. (I wonder why, but good for them!)
(3) No SEC team has yet won a game against a top 25 team, despite 15 such games (no kidding! huh?, wow!)
(4) As a league, the A-10 seems barely relevant.

Very impressive.

I assume that you use Sagarin's composite rating. I prefer his predictor rating. At this point in the season there isn't much difference between the 2. Although there's some shuffling of the order, the top 50 doesn't change, but St. Mary's and Miami drop out of the top 25, replaced by Oregon and South Carolina.

On a side note, Sagarin serves as the basis for RPI Forecast ratings.