Corey's Random Thoughts

Occasional musings too long to fit in my facebook status updates

Tuesday, January 03, 2012

College Football Conference Rankings

With bowl season swinging into high gear, many will be rooting for their favorite conference to prevail.  If you grew up in Alabama, you're almost certainly rooting for the SEC.  And if you grew up in Ohio, you're almost certainly rooting for the Big Ten.  And nothing's going to change your mind.

But what if you've moved around a lot?  Or didn't grow up as a college football fan?  Or attended school out-of-state?  Or attended multiple schools?  For which conference should you root, then?

What follows is a first attempt to rank the conferences by criteria that might be important to fans trying to pick the most virtuous conference.  For the first year of these rankings, I've divided these criteria into three separate categories: academics, football academics, and  football ethics.

Before I discuss the results, I need to stress that these rankings are crude and preliminary.  This is but my first attempt to quantify the merits of each conference.  Due to this fact, I entertained not posting them at all (hence why it's being posted this late into bowl season) . . . but since I already had the tabulations done, I figured I'd post them and note that a lot more could be done to make these better.


Results

The ACC and the Big Ten are the standouts in the academic category, ranking 1-2 in almost category.  The Big Ten is the only conference with 100% AAU membership, but the ACC edges them out for the highest average SAT scores (25th-75th percentile of 1194-1384), highest average ranking in the US News & World Report (49th), and highest graduation rate (83%).  The Pac 12 was a clear third, and the SEC, Big 12, and Big East were clustered at the bottom.  Only 16.7% of SEC schools are AAU members, but the Big East has slightly lower SAT scores (1070-1265), USNWR rank (112th), and graduation rate (67%).

For football academics, I looked at the NCAA APR rating and the graduation rate of the football teams.  The ACC ranked first in both -- 63% of football players graduate.  The Pac-10 ranked at the bottom of the category, but the Big 12 has the lowest APR and the SEC has the lowest football graduation rate -- 51% of football players graduate.

For football ethics, I looked at oversigning and years spent on probation.  The SEC was the biggest offender in recruiting, signing an average of 25.25 recruits per year despite scholarship limits of 85 per team (meaning a large number of recruits have to be encouraged or forced to leave school prematurely or enroll elsewhere).  The Pac-10, however, had the most years spent on probation since 2000 -- an average of 1.83 per team -- in part due to the recent troubles of new conference members Utah and Colorado.  The ACC and Big Ten signed the fewest athletes per year (22.16 and 22.28), while the Big East had zero teams on probation since 2000.

The averages in all the categories I described were standardized so that the school with the highest score had 100 points.  Overall, the ACC ranked first, with 96.41 points out of 100 and the SEC ranked last, with 80.62 points.  This was partially due to the fact that students at ACC schools score about 100 points higher on the SAT than students at SEC schools, and are almost 25% more likely to graduate, but the SEC ranked near the bottom in almost every category.  Without further ado, here are this year's rankings:



Acad Ft. Acad Ft. Ethics Total
ACC 91.67 100.00 97.57 96.41
Big Ten 93.75 95.49 95.92 95.05
Pac 10 79.75 85.62 90.13 85.17
Big East 66.80 88.86 95.94 83.87
Big 12 70.44 82.93 90.63 81.33
SEC 65.23 89.00 87.64 80.62



The full results and tabulations can be found here, and here are details the measures that were used to construct the rankings.  Each of these measures is problematic in multiple ways, but I'll spare you the full discussion (but please note again that these ratings are essentially an imperfect first draft):

AAU Membership
The American Association of Universities is a group of 59 of the most prestigious research universities in the country.  I measured the percentage of schools in each conference that were AAU members.

UWNWR Ranks + Grad Rates + SAT scores
A number of the schools reported ACT scores, which were converted to SAT scores using the concordence tables published by both the ACT and SAT folks (which were identical).  I used the average SAT or converted score for each conference in addition to the average rank and SAT score.

Oversigning
I used the 10 year average of scholarships awarded per year.  Some of these numbers are inflated due to early departures to the NFL and some are deflated due to scholarship restrictions.  But since they're a 10 year average and NFL players still have to remain in school for 3 years, the numbers should be pretty strongly correlated with the number of scholarships being awarded to players who are later removed from the team in one of a number of ways (as detailed by SI).

Academic Progress Rate
This is the NCAA's measure of academic progress -- essentially a way for them to assess whether students on the teams are progressing toward graduation.  I used the latest APR for each school's football program.

Football Graduation Rate
The percentage of football players who graduate within six years.

Years on Probation
I calculated the total number of years spent on probation per team for each conference between 2000 and 2011.


Other possible criteria for future rankings:

Washington Monthly College Rankings
These rankings attempt to rank schools by their service to society rather than their academic prestige.  I entertained the idea of using these rankings and/or pieces of these rankings, but decided against it largely because they were just released and I didn't want to take the time to enter all the data (see what I mean about these rankings being crude and preliminary?)

Sexist Team Names
In a recent study, schools with women's basketball teams who went by different, sexist, nicknames than the men's teams were found to be less likely to hire a female to coach the women's basketball team and gave relatively fewer athletic opportunities for women (as measured by Title IX proportionality).

Arrests
I originally planned on using this data on number of arrests per team, but it only includes last year's to 25 teams . . . and, as this update pointed out, the two programs with the most arrests have had only two arrests between them in the last 6 months -- one of whom wasn't even a practicing member of the team.  If data were available on number of arrests by team, I would certainly use that.

Transfer Policies
This would be hard to measure, but I'd rather not root for teams that prevent their players from transferring closer to home to be near seriously ill parents

If I decide to do another set of rankings in the future, I'm open to any other suggestions as well . . .

Monday, December 05, 2011

Did Rogue Voters Alter the BCS Title Game?

Despite trailing in the computer polls, Alabama beat out Oklahoma State for the second and final slot in the BCS title game by the narrowest margin ever -- averaging 94.19% of the vote across the Harris Poll, Coaches Poll, and Computer Polls to Oklahoma State's 93.33% of the vote (final tally).

The closeness of the vote has led to a closer examination of the human voters.  Sports Illustrated's Andy Staples today writes about a number of renegade voters who didn't rank Oklahoma State either 2nd or 3rd.  SI.com subtitled the piece "Head-scratching ballots played in (sic) key role in deciding who's No. 2."

Did these voters, in fact, wield enough influence to flip the order?  That's actually an easy question to answer.

If we examine the voter breakdown of the coaches' poll, we find that 42 coaches ranked Alabama 2nd and 17 ranked them third.  Meanwhile, 17 coaches ranked Oklahoma State 2nd, 36 ranked them 3rd, 5 ranked them 4th, and 1 ranked them 5th.

Overall, Alabama received 1399 points and Oklahoma St. received 1367 points.  If the 6 voters who ranked Oklahoma State 4th or 5th instead ranked them third, their total would have increased to 1374, a gain of 7 points.  This would have increased their share of the vote from 92.31% of the vote to 93.15% of the vote, a gain of 0.84.

If we examine the voter breakdown of the Harris Poll, we find that 80 voters ranked Alabama 2nd, 33 ranked them 3rd, and 2 ranked them 4th. 35 voters ranked OK St. 2nd, 64 ranked them 3rd, 9 ranked them 4th, 4 ranked them 5th, and 3 ranked them 6th.

In the original poll, Alabama received 2723 points to Oklahoma State's 2654.  If we again assume that only three teams can receive votes, and switch the votes of the 2 voters who ranked Alabama below 3rd and the 16 voters who ranked Oklahoma State below third, we'd again see slightly different results.  In such a scenario, Alabama would now outscore Oklahoma State 2725 (a gain of 2 points) to 2680 (a gain of 26 points).  This would raise Oklahoma State's share of the vote from 92.31% to 93.22%, almost a full percentage point gain.

Under these assumptions, Alabama's score in the computer polls would remain .930, their coaches' poll score would remain .9485, and their Harris Poll score would rise slightly to .9478.  Their new average would be .9421, only .0002 higher than the actual result.

Oklahoma State would gain more points.  Their score in the computer polls would remain .950, but their coaches' poll score would rise to .9315 and their Harris Poll score would rise to .9322.  Their new average would be .9379, up .0046 points from the actual result.

But .9379 would still trail Alabama's .9421.  The difference would be more than halved (from .0086 to .0042), but Alabama would still be playing in the title game.

Did the "Head-scratching ballots" influence the scores?  Yes.  Oklahoma State received fewer points than they otherwise would have.  Did they change the participants in the title game?  No.  Alabama would still be playing LSU even if no voters had ranked either Alabama or Oklahoma State 4th, 5th, or 6th in either poll.

Thursday, August 11, 2011

Super Committee

What are the odds of a "grand bargain" by the so-called "super committee" this fall?  Below is a table of the 12 members and how their lifetime voting records compare to other members of their respective legislative bodies and caucuses.

One could use any number of rankings services or dissect their voting records in various ways and come up with quite different rankings -- as such, this should be interpreted as a first attempt to gain a rough idea of how liberal or conservative the various members are rather than any sort of authoritative study.

That said, the results are somewhat distressing to those hoping for a productive session.  8 of the 12 members are more liberal/conservative than the median member of their caucus and 9 of 12 have voted with their party more than 90% of the time.  Democrat Max Baucus is slightly more conservative than the median Senator and Fred Upton is slightly more liberal than the median Representative, though Rob Portman has also been mentioned as somebody potentially willing to compromise in news reports despite his conservative voting record.

In the Senate it's plain to see that Harry Reid nominated far more moderate members than did Mitch McConnell (by both relative ranks within their party and absolute voting numbers) -- indeed, the most moderate Republican is more conservative than the least moderate Democrat is liberal.  In the House, the story is more muddled.  Relative to their respective caucuses the Republicans are more moderate, but based on their absolute voting records the Democrats could be seen as slightly more moderate.

On average, the Democrats on the committee are more liberal than 49% of their caucus and have voted with their party 89.3% of the time while the Republicans on the committee are more conservative than 63% of their caucus and have voted with their party 95% of the time.  Based on these metrics, the Democrats could be described as more moderate than are the Republicans.

But the political leanings of the average member from each party may be less important than the potential swing votes.  Any chance of the committee's recommendation receiving 7 votes may hinge on Republicans convincing Max Baucus to go along with their plan or Democrats convincing Fred Upton to go along with theirs.  It seems somewhat less likely that the seven most moderate members across party lines would reach an agreement given that only 4 have more moderate voting records than the median member of their party.

It's too early to give up hope entirely, but the voting records of the members of the committee appear to allow little chance of a productive and civil session.  The better idea would have been to allow the four leaders to appoint one or more members from the opposite party to the committee.



Senate DemocratsSenate Republicans
Liberal Rank% Voting with PartyParty PercentileConservative Rank% Voting with PartyParty Percentile
Max Baucus52nd75.49%2%Jon Kyl*16th96.20%66%
Patty Murray*26th91.11%51%Rob Portman12th96.36%74%
John Kerry24th91.50%55%Pat Toomey2nd98.15%96%
House DemocratsHouse Republicans
Liberal Rank% Voting with PartyParty PercentileConservative Rank% Voting with PartyParty Percentile
James Clyburn121st89.48%37%Fred Upton218th84.70%9%
Chris Van Hollen60th93.45%69%Dave Camp150th97.05%37%
Xavier Becerra41st94.49%79%Jeb Hensarling13th97.80%95%



*co-chairs

liberal/conservative rank: 1st corresponds to most liberal/conservative member of that body, the Senate has 53 Democrats (including two Independents) and 47 Republicans while the House has 240 Republicans and 193 Democrats

% Voting with Party: The percentage of the member's vote that have aligned with the party position during their career.

Party Percentile: The percentage of their respective caucus who the member is more liberal (in the case of Democrats) or conservative (in the case of Republicans) than: a higher number represents a less moderate politician.

Sunday, November 28, 2010

Is the SEC the Best Conference?

There are a million ways to answer the question, but the best evidence would be head-to-head match-ups with non-conference teams. Here are the SEC's victories over other BCS conference teams (in order from most to least impressive):

Florida (7-5) beat S. Florida (7-4) at home, 38-14
#8 Arkansas (10-2) won at #19 Texas A&M (9-3), 24-17
#17 Alabama (9-3) beat Penn St. (7-5) at home, 24-0
#11 LSU (10-2) beat W. Virginia (7-4) at home, 20-14
Kentucky (6-6) won at Louisville (6-6), 23-16
#18 S. Carolina (9-3) won at Clemson (6-6), 29-7
Georgia (6-6) beat Georgia Tech (6-6) at home, 42-34
#2 Auburn (12-0) beat Clemson (6-6) at home, 27-24
#11 LSU (10-2) beat N. Carolina (7-5) on a neutral field, 30-24
#17 Alabama (9-3) won at Duke(3-9), 62-13

That the two most impressive victories were against South Florida and Texas A&M really says it all about this list. SEC teams have, collectively, beaten only one ranked non-conference team and two other teams (both from the weak Big East) with fewer than five losses. In short, the SEC's victories against non-conference teams tell us almost nothing about the strength of the SEC.

What about the non-conference losses? Here are the SEC's losses to other BCS conference teams (in order from least to most embarrassing):

Vanderbilt (2-10) lost at home to Northwestern (7-5), 23-21
Tennessee (6-6) lost at home to #1 Oregon (11-0), 48-13
Vanderbilt (2-10) lost at UConn (7-4), 40-21
Florida (7-5) lost at #22 Florida St. (9-3), 31-7
Vanderbilt (2-10) lost at home to Wake Forest (3-9), 34-13
Georgia (6-6) lost at Colorado (5-7), 29-27

The list doesn't include Ole Miss's loss to Jacksonville St. and also doesn't shed a whole lot of light on the SEC's strength, as most of the match-ups were lopsided or involved Vanderbilt.

Overall, the SEC was 10-6 against BCS conference teams and 1-2 against ranked BCS conference teams. The conference was 1-5 when playing teams with fewer losses; 2-0 when playing teams with identical records; and 7-1 when playing teams with worse records. The conference was 5-3 at home; 4-3 on the road; and 1-0 on a neutral field.

So, what does this tell us about the strength of the SEC? Not much. 16 games played by a conference of 12 teams is almost never going to tell us all that much. It seems obvious both that the SEC isn't far worse than every other conference but also that it isn't far better than every other conference this year. I'm not sure anybody would disagree that the SEC has been the strongest conference over the past five years, but we simply don't have enough evidence to conclusively claim the same this year.

Sunday, October 24, 2010

Top 25 Update

How things appear after today's games, with strengths and weaknesses of each team:

teams that look really good so far

1.) Oregon
strengths: outscored Stanford 49-10 the last three quarters; dominant in 2nd half all season; explosive offense
weaknesses: outscored in the first half by Arizona St.; let Washington St. and Tennessee hang around too long

2.) Boise St.
strengths: beat VaTech; handled Oregon St.
weaknesses: no decisive victories over top teams; weak schedule

3.) TCU
strengths: beat Oregon St.; crushed Baylor; crushed Air Force
weaknesses: were only up 6-0 at halftime vs. Colorado St.


teams that might be really good

4.) Auburn
strengths: beat South Carolina; beat Arkansas; the next Tim Tebow at QB

weaknesses: almost lost to Mississippi St.; almost lost to Kentucky; almost lost to Clemson

5.) Alabama
strengths: crushed Florida; beat Arkansas; dominant running game
weaknesses: lost to S. Carolina; inconsistent passing game; young defense

6.) Michigan St.
strengths: beat Wisconsin; beat Michigan
weaknesses: needed a trick play in OT to beat Notre Dame; barely beat Northwestern

7.) Arizona
strengths: soundly beat Iowa; crushed Washington
weaknesses: lost to Oregon St.

8.) Missouri
strengths: beat Oklahoma
weaknesses: no other victories over top teams

9.) Oklahoma
strengths: destroyed Florida St.; beat Air Force; beat Texas
weaknesses: barely escaped Cincy

10.) Wisconsin
strengths: beat Ohio St.; beat Iowa
weaknesses: struggled far too often against inferior teams

11.) Ohio St.
strengths: handled Miami; dynamic QB
weaknesses: escaped Illinois; no victories over top teams

12.) Stanford
strengths: beat USC; destroyed UCLA; strong-armed QB
weaknesses: outscored 49-10 by Oregon the last three quarters

13.) Utah
strengths: beat Pitt; blew out Iowa St.
weaknesses: only beat Pitt by 3 in OT; have yet to play a top team

14.) Florida State
strengths: beat Miami soundly; blew out 4 of other 6 opponents
weaknesses: blown out by Oklahoma; barely beat BC


teams that might be pretty good

15.) South Carolina
strengths: beat 'bama
weaknesses: lost to Kentucky; struggled to put away Vandy 

16.) Nebraska
strengths: beat Oklahoma St.; destroyed Washington; destroyed Kansas St
weaknesses: lost to Texas; only beat South Dakota St. 17-3

17.) Iowa
strengths: soundly beat Penn St.; beat Michigan; elite defense
weaknesses: soundly beaten by Arizona; no victories over top teams

18.) LSU
strengths: beat Florida; played Auburn close; luckiest coach on Earth
weaknesses: almost blew a huge lead against a depleted UNC team; should've lost to Tennessee at home; almost lost to W. Virginia at home; took a 10-3 lead against Vandy into the 4th quarter; inconsistent offense


teams that haven't proven much of anything

19.) USC
strengths: crushed Cal; almost beat Stanford
weaknesses: lost to Washington; no victories over top teams

20.) Virginia Tech
strengths: almost beat Boise St.; beat NC State
weaknesses: lost to James Madison

21.) Oklahoma St.
strengths: undefeated; high-scoring offense
weaknesses: no impressive victories; almost lost to Texas A&M; gave up 51 points to Nebraska

22.) Arkansas
strengths: almost beat Alabama; strong-armed QB
weaknesses: barely beat Texas A&M; gave up 65 points to Auburn; unimpressive win vs. Ole Miss; interception-prone QB

23.) Miami
strengths: crushed Pitt; crushed UNC
weaknesses: crushed by Florida St.; handled by Ohio St.; interception-prone QB

24.) Mississippi State
strengths: beat Florida
weaknesses: barely beat UAB;

25.) Nevada
strengths: destroyed Cal; prolific offense
weaknesses: struggled to put away BYU; lost to Hawaii


teams with resumes no less impressive


NC State
Pitt
Oregon St.
Baylor

Sunday, October 17, 2010

Top 25 Update

How things appear after today's games, with strengths and weaknesses of each team:

teams that look really good so far

1.) Oregon
strengths: outscored Stanford 49-10 the last three quarters; dominant in 2nd half all season; explosive offense
weaknesses: outscored in the first half by Arizona St.; let Washington St. and Tennessee hang around too long

2.) Boise St.
strengths: handled Oregon St.
weaknesses: escaped VaTech

3.) TCU
strengths: beat Oregon St.; crushed Baylor
weaknesses: were only up 6-0 at halftime vs. Colorado St.

4.) Oklahoma
strengths: destroyed Florida St.; beat Air Force; beat Texas
weaknesses: barely escaped Cincy


teams that might be really good

5.) Alabama
strengths: crushed Florida; beat Arkansas; dominant running game;
weaknesses: lost to S. Carolina; inconsistent passing game; young defense


6.) Auburn
strengths: beat South Carolina; beat Arkansas; the next Tim Tebow
weaknesses: almost lost to Mississippi St.; almost lost to Kentucky; almost lost to Clemson

7.) Michigan St.
strengths: beat Wisconsin; beat Michigan
weaknesses: needed a trick play in OT to beat Notre Dame

8.) Arizona
strengths: soundly beat Iowa
weaknesses: lost to Oregon St.

9.) Iowa
strengths: soundly beat Penn St.; beat Michigan; elite defense
weaknesses: soundly beaten by Arizona; no victories over top teams

10.) Wisconsin
strengths: beat Ohio St.
weaknesses: struggled far too often against inferior teams



11.) Ohio St.
strengths: handled Miami; dynamic QB
weaknesses: escaped Illinois; no victories over top teams

12.) Stanford
strengths: beat USC; destroyed UCLA; strong-armed QB
weaknesses: outscored 49-10 by Oregon the last three quarters

13.) Utah
strengths: beat Pitt
weaknesses: only beat Pitt by 3 in OT; have yet to play a top team

14.) Florida State
strengths: beat Miami soundly; blew out 4 of other 6 opponents
weaknesses: blown out by Oklahoma; barely beat BC


teams that might be pretty good

15.) South Carolina
strengths: beat 'bama
weaknesses: lost to Kentucky; inconsistent offense 



16.) Nebraska
strengths: destroyed Washington; destroyed Kansas St.; talented QB
weaknesses: lost to Texas; only beat South Dakota St. 17-3; young QB

17.) LSU
strengths: beat Florida; luckiest coach on Earth
weaknesses: almost blew a huge lead against a depleted UNC team; should've lost to Tennessee at home; almost lost to W. Virginia at home; took a 10-3 lead against Vandy into the 4th quarter; inconsistent offense; playing poorly and using poor strategy is bound to backfire eventually

18.) Texas
strengths: beat Nebraska; loads of talent
weaknesses: blown out at home by UCLA; no impressive victories; inexperienced QB




teams that haven't proven much of anything

19.) Oklahoma St.
strengths: undefeated; high-scoring offense
weaknesses: no impressive victories; almost lost to Texas A&M


20.) Arkansas
strengths: almost beat Alabama; strong-armed QB
weaknesses: almost lost to Georgia; barely beat Texas A&M; interception-prone QB; suspect defense

21.) West Virginia
strengths: 5-1
weaknesses: no really impressive victories; almost lost to Marshall


22.) USC
strengths: crushed Cal; almost beat Stanford
weaknesses: lost to Washington; no victories over top teams

23.) Virginia Tech
strengths: almost beat Boise St.; beat NC State
weaknesses: lost to James Madison

24.) Missouri
strengths: undefeated; beat San Diego St.

weaknesses: haven't beaten any other teams with a winning record

25.) Nevada
strengths: destroyed Cal; prolific offense
weaknesses: struggled to put away BYU; lost to Hawaii

teams with resumes no less impressive




Mississippi St.
Miami

Kansas St.
North Carolina
NC State
Rutgers
Oregon St.

About Me

Lewiston, New York, United States