INDIANAPOLIS -- Ever wonder how the NCAA's men's basketball committee weighs the differences between a hotshot, mid-major power (hello, Butler) and a good team from a great league (Marquette)?
Ever question why Duke or North Carolina always seem to play tournament games in Charlotte or Greensboro?
And just how do you decide that Vermont is a 15 seed in the East, Northern Iowa is a 13 in the South and Oral Roberts is a 14 out West?
I've often pondered these questions over the last 20 years. So when the NCAA invited me to its fourth annual mock bracket session at its headquarters in Indianapolis, I jumped at the chance. Twenty members of the media, along with some conference officials, received an inside-the-war-room look at just what goes into selecting, seeding and bracketing the field of 65 teams.
The "real" committee of 10 members is now monitoring teams, watching games and getting ready to arrive at the Westin Hotel in Indy on the Wednesday before Selection Sunday (set for March 14 this year). Then they'll begin a process that can include a dizzying amount of numbers, enough computer screens to make you cross-eyed, plenty of drinks and snacks and basketball arguments that would make Archie Bunker proud.
By the time CBS reveals the group's work on Sunday evening, it's all but impossible for even the very best committee members not to be completely fried.
"We had one year where we wanted to call CBS and tell them, sorry, we'll have to get back to you with that bracket," said Greg Shaheen, the NCAA's senior vice president of basketball and business strategies who coordinates the tournament and its selection weekend. "We just had so many teams under consideration. But that Sunday deadline is always hanging out there and you work towards it."
The five days of work at the Westin was melted down to 15 hours, spread over two days, for the mock exercise. Everyone worked off a "pretend" end to the regular season as of last Wednesday. We were paired with another attendee (thanks to Will Roleson of the Horizon League) and sat in front of a laptop and a bank of three screens that were always flashing updated "nitty-gritty" sheets filled with a team's schedule and various statistics and others that tracked the teams in the field or under consideration.
Will and I took the place of an actual committee member, Wake Forest athletic director Ron Wellman. A committee member can't vote on his own school or conference and if his team is ever a topic of discussion, he must leave the room. Wake, fortunately, was an easy addition to the field so we didn't get much exercise.
On the first ballot, the committee votes on teams it feels should be in the field with no questions asked. You need eight of 10 votes to make this cut and 19 teams (such as no-brainers like Kansas, Syracuse and Kentucky) were quickly in. Next came generating the group of teams that would be under consideration for the remaining at-large spots (there are 34). Teams need at least two votes to reach this pool and we offered 53, a number that prompted Shaheen to jokingly say "so you all want to be here until March, I suppose."
To move ahead with selections, the group was asked to select the next best eight teams in the under consideration pool. After a consensus was reached on those eight, the voters were directed to list the eight in order. The top four of that vote were added to the field. That process, which Shaheen termed "slicing the onion," is then continuously repeated until the field of at-large teams is filled out.
At any point, a member can call out to NCAA staffers for a team sheet on a school which are instantly projected on the screens at everyone's station, often two and three at a time. Included in the sheets are basics like win-loss record and RPI but also a very illustrative breakdown of how you fared against teams with RPI's between 1-50, 51-100, 101-200 and 201 and above. The computers also break down your average RPI win and loss, further crystallizing exactly who you played and who you beat.
Of the eight finalists on the first vote, four teams (Baylor, Georgia Tech, Texas A&M and Wake Forest) successfully emerged. Others were thrown back into the pot and again identified in a final eight. This time Rhode Island, Richmond, Northern Iowa and UNLV emerged in the field.
The committee always insists that RPI is just one tool in the process. That's not really true. The RPI is the dominant weighing system the NCAA uses to compare teams, basically because there is no other fair one out there. While no one in our committee openly said, "yeah, but California's RPI is better than South Florida's," most of the numbers on a team sheet are RPI-based. Even so, it was flip-a-coin tough to differentiate between two worthy candidates. To our credit, big wins and bad losses (vs. the RPI, of course) seemed to be the most important calling card.
At one point, Xavier (RPI of 27) was matched up against Illinois (72) in a two-team tiebreak vote. When the team sheets were side-by-side it became apparent that Illinois' recent wins over highly rated Wisconsin and Michigan State trumped anything the Musketeers had accomplished. The Illini won that vote.
Before the at-large field was filled, we started seeding teams with a goal of compiling a seed list from 1-65. The seeding is done the same as selection by selecting eight teams and then ranking those results. The first stab revealed that Kansas was the overall top seed, followed by Syracuse, Kentucky and Villanova. Next came Purdue, Georgetown, Duke and West Virginia, giving the Big East four of the top eight seeds.
We returned to selecting to end Thursday's late session and added Florida State, Marquette, Ole Miss, Missouri and Xavier, bringing the at-large total to 36. There are only 34 spots but because some of the at-large teams will make it into the tournament as automatic qualifiers via their conference tournaments, we know we'll need more teams.
We returned the next morning armed with fresh results from conference tournaments. Those results produced only one game of consequence we had to wait on the ACC final between Duke and Virginia. If the Cavs won, our someone could lose an at-large spot.
That set up a final, final vote. There were 25 teams left on the board, so we voted for the top eight and then listed the results in order. Those results were razor-thin close with St. Mary's, Charlotte and South Florida getting in. If Duke beats Virginia, Florida would sneak in.
Armed with the full field, we ripped through the seeding and formed the complete seed list. Then it was on to bracketing. Shaheen says this normally takes the committee about an hour, and the NCAA has loads of important bracket rules, like placing teams into the field 1-to-65 with an eye toward keeping everyone as close to home as possible, putting the first three teams in a conference in different regions and trying not to put the top-five seeds at a crowd disadvantage in the first round.
Some observations: It is almost impossible for a committee member to keep a running total of teams from a single conference in his head. Our committee did not put any weight into so-called "good losses." Shaheen says numerous committee members over the years have strongly advocated for one particular team or another, even if their numbers didn't look so hot. This is a sport, after all, not a math class.
Bottom line? Each committee member favors different criteria. Slicing the onion for the last five teams or so is next-to-impossible, and you can easily be out-voted even if you feel passionately about a team's chances. It's a hard job, but one that any basketball-crazed fan would love to own.