Tuesday, April 1, 2014

PAC-12: the Most Under-Valued Conference in the Country?

Update:  I've updated the stats to include the Final Four.

Brief Observation
I'm impressed by Kentucky's deep run.  I don't think anyone can beat them they way they are playing.  But I also think that if Michigan had not lost center Mitch McGary this season, they would have flattened everyone. That is one talented, well-coached team.


The Numbers Don't Lie
I thought I'd post my conference seeding-versus-performance analysis while there's still interest prior to the finals. 

My methodology is simple:  I track how teams do in the NCAA Tourney relative to their seeding, then aggregate the results by conference.  So a #2 seed that loses to a #5 seed gets a minus 3, while the winner gets a plus 3.  Winning/losing as your seed says you should expect earns a zero.  So if your conference is around zero, plus or minus a few, you're doing pretty much as expected.  Keep in mind, this doesn't tell you whether a conference is good or bad.  It only indicates whether expectations as assigned by the seedings were warranted.

I've also included an additional data point, that of how many teams in each conference over-achieved, under-achieved or did about as expected.  This helps tell us whether or not the aggregate number is attributable to conference performance as a whole or to a single data point (i.e., a major upset). 

The lopsided seeding match ups for the Final Four mean that the SEC can move even higher or drop, the AAC can still move up, and the best the B1G can do is stay roughly where they are or, more likely, drop even further.     (Update: the SEC remained static, the AAC rose and the B1G sank further.)
  2014                      # Teams    # Over Achievers/# Neutral/# Under Achievers

SEC              +22           3                 2 / 0 /1
PAC 12           +8           6                 1 / 3 / 2
AAC               +6           4                 1 / 1 / 2
A-10              + 2           6                 1 / 3 / 2
MWC              -3           2                 0 / 1 / 1
Big East          - 8           4                 0 / 2 / 2
B1G               -15           6                 2 / 1 / 3
Big 12            - 16          7                 1 / 3 / 3
ACC               -21          6                 1 / 2 / 3
There are two big changes from last year.  First, the SEC skyrocketed, owing to Kentucky's deep run.  The ACC, on the other hand, plummeted.  Blame Duke and Syracuse (with some help from Dayton.)

Here are aggregate numbers for four of the last five years (I just didn't get around to it in 2012):

By Year     2014      2013     2011   2010    Aggregate
PAC 12          +8       +17       +5       +13        +43
SEC             +22         -8        -8         -6             0
ACC             -21         -2        +4          0          -19
B1G              -15         -8       -15        -11         -49
Big 12          - 16        -16       -14       -15         -61
Big East         - 8        -13       -28       -35          -84

Uh, does anyone see a pattern here?  One year is perhaps a statistical outlier.  Two years, maybe something forming.  But this?

It's not even close.

I would expect for most big conferences to trend slightly negative, owing to large contingents in the tourney (except for the SEC, which has been sending three teams for a few years), and the occasional success of mid-majors. Interpreted as such, the B1G benefits from a significant and consistent favorable bias.

However, the PAC 12 (undervalued), the Big 12 and the Big East (both severely overvalued) are clearly off the charts relative to one another (although the latter got some relief recently as it fragmented and took some of the bias with it).

The numbers don't lie.