Debunking Popular NCAA Tournament Selection Myths

This edition of Cardinal basketball doesn't fit the mold of any recent teams, with such big swings up and down. Beat Florida and Xavier, but drop to Montana and Washington. So how to forecast where the team may be seeded come March? Navigating that process is frought with many myths, and we are here to separate the truth from fiction in the NCAA Tournament selection process.

College basketball fans tend to get hung up on the number of bids they think a conference should get based on the strength of the conference overall or, worse yet, based on the number of teams that conference has historically gotten into the dance.  Despite the prevalence of the myth that the NCAA Basketball Committee utilizes conference strength and the historical number of bids a conference has received in determining at-large tournament bids, the selection process doesn't work that way at all.  As we near the halfway point of conference play, and fans are increasingly speculating about tournament bids and seedings, it seems to be an appropriate time to dispel some popular myths.

The best starting point for any discussion about the NCAA tournament selection process is with the NCAA's published principles and procedures for selecting and seeding teams for the tournament.  Here's the link:

http://www.ncaasports.com/0,5920,1_1661_0_24453,00.html#select

Perhaps the most popular myth is that the strength of a conference determines the number of bids its member teams receive.  For example, there has been considerable speculation that because the Pac-10 is having a down year overall, as measured by conference RPI and other "objective" measuring tools (as well as subjective observations), the conference will receive only three (or maybe four) bids.  It's true that the Pac-10 is having a poor year overall by any metric.  However, it would be a mistake to assume that it necessarily follows that the Pac-10 will be limited to a certain number of bids as a result. 

Note the first principle for selecting at-large teams:  "The committee shall select the best available teams to fill the at-large berths, regardless of conference affiliation."  History demonstrates that the Committee has followed this principle faithfully and has evaluated teams on their individual merits without regard to conference affiliation. 

A result of the Committee's approach to selecting teams based on their individual merits without regard to conference affiliation is that it's impossible to guage the number of bids a conference will receive without a careful examination of the postseason resumes of the individual teams in that conference.  It's entirely possible for a strong conference to garner a smaller number of bids than a relatively weaker conference, or for two conferences of comparable strength to receive disparate numbers of bids. 

A couple of examples jump to mind.  In both the 1999 and 2000 seasons, the ACC only got three teams in the dance despite being one of the top conferences.  The fundamental problem for the ACC each of those seasons was that while the top three teams were often dominant, all the other teams finished with not only mediocre to poor conference records, but poor overall records.  It didn't help that teams that finished in the middle of the conference standings had weak out of conference schedules.  The result was that each of the teams outside of the top three didn't merit selection based on the Committee's criteria.  For example, in 2000, Virginia, which finished fourth and beat North Carolina twice, had a strength of schedule that wasn't in the top 100 thanks to an out of conference schedule filled with cupcakes, and its RPI and other factors did not warrant a bid.  Coach K whined about the ACC's only receiving three bids for the second straight season, but the fact was that according the Committee's stated criteria, only three ACC teams merited bids when the teams were evaluated on an individual basis without regard to conference affiliation.

On the other hand, a strong conference with more balance than the ACC had in the 1998-99 and 1999-00 seasons has the opportunity to garner many more bids.  For example, seven Big Ten schools made it to the dance in both 1999 and 2001.  In each of those seasons the Big Ten standings featured a large number of teams at the .500 level or slightly better in conference.  Most Big Ten teams those years played tough non-conference schedules, and the conference as a whole faired reasonably well in those out of conference games.  As a result, the Big Ten had seven teams in the '99 and '01 seasons that had impressive resumes, according to the Committee's stated criteria. 

If one were to evaluate on any subjective or objective basis the strength of the ACC conference as a whole versus the strength of the Big Ten conference as a whole in the 1998-99 season, it would be difficult if not impossible to conclude that the Big 10 was so much better than the ACC that it deserved 7 bids to the ACC's 3 bids.  But again, the Committee doesn't select at large teams based on conference strength.  The selection of the at-large teams in 1999 is proof positive that the Committee is faithful to its duty to evaluate teams on an individual basis and to ignore conference affiliation.  The ACC carries as much influence as any conference in the country, yet for all its lobbying, it couldn't persuade the Committee to deviate from the selection principles mandated by the NCAA.

The examples above illustrate another important principle.  The number of bids a conference has historically received is irrelevant to the Committee's selection process.  The ACC has received as few as three bids, but as many as six.  The number of bids received by the Big Ten and Big East also fluctuates wildly from year to year.  Pac-10 fans know that the conference has received anywhere from three to six bids in recent years.  The success of a conference in previous years' tournaments also has no bearing on the selection process.  The Big Ten has had some spectacular flameouts in years past, but that didn't prevent the Committee from selecting as many as seven teams in the years immediately following those flameouts.

With the Committee's selection principles in mind, let's turn to the Pac-10.  Notwithstanding the fact that the conference was much stronger overall last season, the Pac-10's at-large bids may prove instructive for an analysis of the prospects for the conference this season.  No team dominated the conference last year, with the result that six teams were able to finish with 11-7 records or better.  Those six teams played sufficiently well out of conference that each had a strong post-season resume and merited a bid according to the committee's selection criteria.  When all the nervous nellies were fretting about the number of bids the conference had historically received, those familiar with the selection process knew not to be concerned that the conference would be limited to some specific number of bids based on the conference's overall strength or the historical number of bids received, which had typically ranged from three to five.

When the Committee meets this March, it's going to evaluate all teams, including all Pac-10 teams, on their individual merits, without regard to conference affiliation.  So while it's easy to assess the Pac-10's overall strength at this point, given that there are only a few out of conference games remaining, it's impossible to know how many at-large bids Pac-10 teams might receive.  It's entirely possible that anywhere from three to five Pac-10 teams could get bids, depending on how the remainder of the season plays out.  I'll offer up two examples, neither of which is implausible, to illustrate the range of possibilities.  Suppose the top half of the standings wind up looking like this:

Arizona (17-1)
Cal (13-5)
Stanford (12-6)
Oregon (10-8)
Arizona St. (9-9)

In a scenario like that, there's a good chance that only the top three teams would get bids (although the Pac-10 tournament would offer bubble teams an opportunity for additional wins, perhaps quality wins).  It's unlikely that Oregon and/or ASU would have a decent case for an at-large bid based on the Committee's stated selection criteria.  For example, their overall records and SOS wouldn't be strong enough, and neither team has much in the way of quality out of conference wins. 

Let's take another example:

Arizona (17-1)
Arizona St. (12-6)
Cal (12-6)
Stanford (11-7)
Oregon (11-7)

In that scenario, five bids would be a distinct possibility.  (I deliberately picked a scenario in which the fourth place teams would almost certainly have a stronger case for a bid than the third place team, based on strength of schedule, quality wins, RPI, etc.)  Of course, the standings in my second example suggest a conference in which there is a substantial gap between the haves and have-nots in the conference.  However, last year's standing demonstrate not only that such a bifurcated conference is possible, but that in such a scenario, the number of possible bids is probably greater than in a scenario in which many teams are bunched at around the .500 mark. 

A couple of additional points bear mention with regard to the number of bids a conference is likely to receive.  I'm not suggesting here that a correlation doesn't exist between overall conference strength and the number of at-large bids a conference receives.  Of course there is a tendency for better conferences to receive more bids, but that's only because the individual teams in stronger conferences tend to have better overall records, greater strengths of schedule, more quality wins, higher RPIs, etc. (i.e. all the things the Committee considers in the selection process.)  Although the statistical correlation I referred to above exists over a large enough sample size, in predicting the number of bids a conference is likely to receive in a given year, one must analyze the individual teams in the conference utilizing the Committee's stated selection criteria.  It's simply a mistake to let a conference's overall strength or weakness color the analysis of the individual teams. 

For anyone wanting a better understanding of how the RPI is calculated and used by the Committee, I would suggest reading the NCAA release available on the internet at:

http://www.ncaa.org/news/1999/19990215/active/3604n35.html

The best approximation of the actual RPI used by the Committee is available at Jerry Palm's website at www.collegerpi.com.  Palm's explanation of how the RPI is calculated and how the Committee hides the ball is excellent and worth reading. 


The Bootleg Top Stories