Monday, March 14, 2011
How I Learned to Stop Worrying and Love the Selection Committee
The NCAA Basketball Selection Committee is in many ways like the Supreme Court. Its decisions are done behind closed doors and many are outraged after the verdicts are reached. After those decisions are made, many suspect some sort of insidious motives or personal agendas. Sometimes the question is posed if the committee members/justices have their heads on straight. Even the numbers on each panel are similar, nine for the Supreme Court and 10 for the committee.
There is one big difference, though. As an institution, the Supreme Court almost always registers approval ratings above 50 percent. One would suspect that the majority of college basketball fans do not have approbation of the selection committee in any year.
After Sunday's Selection Show for the NCAA tournament, the most surprising edition of selections since 2007 at least, that low level of trust was well and alive. Seth Davis and Greg Anthony were puzzled on CBS. Jay Bilas asked if "the committee knew if the ball was round." Digger Phelps insinuated that there was some latent non-BCS at-large quota the committee had to meet after expanding the tournament to 68 teams and 37 at-large bids.
I choose to believe NCAA basketball supremo Greg Shaheen, who every year brings in a group of media personalities to participate in a mock selection committee and runs it the exact same way that the committee operates. Does he have incentive to lie to the media if somehow the committee has institutionalized a "judicial activism" as part of its normal affairs? I suppose he has some, but there have been far too many people serving on the committee over the years with a wide array of conference and school affiliations for a secret agenda to be feasible and stay under wraps. Furthermore, the committee has not been consistent enough on issues like a mid-major quota over the years for that agenda to be at work.
What the committee ultimately decides on is inherently rooted in the principles of groupthink. Out of a 10-member committee, six have to agree on something for it to get voted through. This is why 88 individuals projecting the process is a foolish endeavor. Different groups of people can come up with wildly divergent preferences, especially if a compelling argument on behalf of a certain viewpoint takes root. Also, different committee members may value different factors of a team's resume to varying extents. Such is the existence of a subjective committee. The alternative would be computers selecting the teams, something that has been mentioned by no one as a practical substitute for good reason. A compromise between the two approaches would be to select teams in a similar vein to the oh-so-popular BCS.
On a personal note, I used to be one of these individuals who tried to predict the decisions of the 10-member group. Three out of the four years I posted a mock bracket on the Internet, it was rated better than Joe Lunardi's final submissions, based on a formula that includes the correct number of teams selected, as well as the numbers of teams within one seed line of their actual position. Last year, I was outraged when Florida got in over Illinois, who I thought had a superior case. Florida subsequently gave BYU a heck of a game. In the end, it became ridiculous to try to replicate something that is only truly replicable with an abundance of time and capable people with the right knowledge.
This year, the inclusions causing the biggest uproar are UAB and VCU, with Clemson and USC providing an auxiliary level of grumblings. Those four teams provide the at-large component of the First Four in Dayton. Three of the four would not have been in the field if last year's format were in place. In a year in which everybody seemed to cite the existence of a weak, wide-open bubble, the outrage over teams being "snubbed" seems curious. UAB and VCU have not put together resumes that would be confused with a middling Big West or MEAC team on any day of the week.
UAB won the regular season conference championship of a top-10 league outright and went 10-7 against the RPI Top 100. Many teams in the field can't cite a winning record against the upper third of Division I. The Blazers possess an excellent point guard in Aaron Johnson. Forward Cameron Moore and Guard Jamarr Sanders provide an inside-outside scoring combo. UAB has C-USA's best defense not named Memphis.
VCU beat the CAA's other two NCAA representatives, Old Dominion and George Mason. The Rams also took care of UCLA on a neutral floor, and won at Wichita State, one of the most historically impenetrable venues outside of the BCS conferences. Four players average in double figures, and VCU ranks only behind a superb George Mason team in offensive efficiency among CAA teams.
If I were to singularly pick NCAA teams, would I have chosen UAB or VCU? Probably not. Yet, one person picking teams in a vacuum is not how the process works.
The fact remains that, with the exception of the Ivy League, the only way to be completely sure about your fate for the tournament is to win your league tournament. Short of that, you have to win at a regular clip against quality teams and not suffer bad losses to be reasonably confident about where you stand on the second Sunday of March. Colorado and Virginia Tech did not meet either of those requirements, each suffering three bad losses. The Buffaloes and Hokies lost 13 and 11 times overall, respectively.
This is not to say that the selection committee does everything right. Its lack of transparency leaves a lot to be desired and leaves it open to even educated people formulating extravagant conspiracy theories about it. If we can rule out the committee room ever having cameras with live audio available on a pay-per-view channel, the NCAA should at least make the transcripts of the committee proceedings available. The NCAA could probably make a lot of money off the transcripts, if it were to make each day's proceedings available for download at a $10 or $20 cost at 10 PM ET the week or so it is in session.
Additionally, the RPI has become a rump metric, if it wasn't already before. It is three-quarters based on some form of schedule-strength and only takes wins and losses into account. A new formula incorporating things like margin of victory, offensive efficiency and defensive efficiency needs to arise. Committee members and those who have participated in mock selections will tell you that a team's RPI number is not important, but the baselines by which schedule strength and quality wins are measured are still based on RPI.
Every year, once tournament games begin, the debate about who belonged in the field seems to melt away. After all, it's extraordinarily unlikely that Virginia Tech, Colorado, or Saint Mary's would have won the national championship were any or all to have been included.
Even assuming that the committee has messed up two or three selections this year, and that left-out teams were indeed of higher quality than those included, that still means the committee has picked about 93% of the most deserving at-large teams. I dare even the most passionate of BCS supporters to argue that an aggregation of the polls and computer ratings has allowed that system to get the teams playing for the championship in college football right 93% of the time.
If the selection process were to solely use computer rankings to decide the field, the problem would be worse. If at-large teams were picked based on the highly respected Pomeroy ratings, Maryland, Nebraska, and New Mexico would all be on their way to the NCAA tournament.
The selection committee is not perfect by any means. It can most certainly be improved in future years by finding something to replace RPI. However, its existence is better than any of the alternatives available for selecting teams to participate in the tournament. The use of the selection committee by the NCAA in its current composition is something that should be largely celebrated, not condemned.