When the NCAA tournament selection committee hands in its 68-team bracket early Sunday evening, the committee will have finished what is essentially a season-long research project.
Selection Sunday is the final exam; the committee builds the bracket in just one day.
Until several years ago, what went on behind those closed doors over five days in Indianapolis was as secretive as the CIA. The 10 committee members are still sworn to secrecy, and their votes are not made public; however, one man has made sure we understand how they get to 68.
He is Greg Shaheen, who for 12 years oversaw the process as the executive vice president for championships and alliances at the NCAA.
“I was a big fan of transparency over my time at the NCAA and made significant strides in that regard,” Shaheen told Bleacher Report.
Shaheen is no longer with the NCAA and will not be in the room on Sunday for the first time since 2000, but that won't stop him from his continued mission to help us understand how the committee gets from 347 to 68.
How the Committee Selects the Field
Let’s go ahead and get this out of the way.
If there’s a flaw to the NCAA’s selection process, it’s that the committee’s crutch is something called the Ratings Percentage Index, known as the RPI.
The NCAA uses the RPI, invented in 1981, as a ranking system in 11 different men’s and women’s sports. The rankings are based on winning percentage against Division I opponents (25 percent), opponents’ success (50 percent) and the strength of schedule of the opponents of your opponents (25 percent).
The RPI feels archaic in college basketball, where new basketball-specific rating systems have become mainstream, such as Ken Pomeroy or Jeff Sagarin’s ratings. Those systems not only consider strength of schedule but also how teams perform statistically—delving deeper than simply wins and losses.
As I would often say in the room, there would be discussions about maybe doing away with the RPI, my comment would always be, "That’s perfectly fine. What do we want to replace it with? You want to replace it with overall record; you want to replace it with home record, road record?"
The bottom line is you have to organize the information somehow, and that’s what our key learning was, you have to organize it somehow. That’s the primary purpose of what the RPI is used for in the room.
This is one misconception that Shaheen wants to clear up: The committee doesn't select teams strictly by RPI.
That process begins on Wednesday afternoon when each member reports on his or her conferences. The 31 conferences are split between the members before the season, and each committee member receives the satellite packages that allow him or her to view every game
Committee members also check in with representatives from each conference, receiving injury updates or any other relevant information.
After reports are given on Wednesday afternoon, each member submits an initial ballot for the 37 at-large teams that will be selected. In the first column, each member lists the teams considered no-brainers that require no further debate. Whether that is 20 teams, 25 or another number at that point is up to the committee member.
The teams that receive all but two eligible votes are put in the field. If a member is the athletic director at a school or commissioner of a league, he or she is not eligible to vote for those teams.
The second column on the initial ballot is for any teams that deserve to be considered and debated, and the teams that receive three or more mentions are thrown into the pool.
“There can be any number of those, although as we learned more about the process, we really tried to make sure they didn’t load it down,” Shaheen said.
Into the weekend, the committee continues to whittle down these teams until it can agree on 37. Conference tournaments can influence this, as some of the initial at-larges will become automatic qualifiers.
Other than the results of the conference tournament, Shaheen insists it does not matter which conference a team plays in. While bracketologists might say Conference X will receive this many bids, the committee does not concern itself with that.
“You’re picking them as independents, primarily,” Shaheen said. “It has nothing to do with conference affiliation.”
This is where the real debate begins.
The committee compares teams side by side, and the results of conference tournaments can move teams up and down the board. Each member has his or her own rankings, and eventually the ballots will decide the overall seeding.
“They analyze head-to-head matchups, common opponents, really dig into box scores,” Shaheen said.
The RPI can help, as each team’s resume—or “nitty gritty” report—is available at the click of a button. As Shaheen said, these reports are not meant to be the end-all criteria; they are used more to organize the information.
For those loathers of the RPI, Shaheen worked during his time to provide committee members with any other information they wanted, even if it was other rating systems.
Back in the day, that information wasn’t widely distributed in the room, but over the last 12 years that I was there, we really worked to make it more and more accessible. Because everybody has to work to develop their own style and approach.
And there’s no need to make it more complicated. There’s no need to act like those measures don’t exist. We might as well have them available if committee members want them.
Seeding is the process that takes up the most time during the five days, and imagine doing it without technology. It used to be that reports for each team were printed out for each committee member.
My first year in the room was 2001 and there was nothing in the room that year. No computers of any kind. So there literally were tens of thousands of photo copies made throughout the weekend. ...
The balloting was all done on notepad sheets of paper. So you’re dealing with handwriting and who knows what half these people are writing.
During that first year, Shaheen brought his laptop in and used an analog dial-up line to get score updates. That offseason he had the NCAA’s IT department start a bracketing system and set up the room with computer monitors throughout.
Televisions are not allowed in the room, so committee members come and go so they can watch the conference tourney games that may influence some of their decisions.
Once the committee is ready to turn in its final rankings from one to 68, the bracketing begins.
Shaheen said that most believe the committee fills out the bracket throughout the five days in Indy, much like the bracketologists do it; however, teams don't slide into any slots until Sunday.
That comes with a caveat, as the bracket does start to take shape at the top and the bottom throughout the week.
One group, identified as the “first quadrant group,” meets during off hours to figure out how to rank and place the top 16 teams in the bracket. The “fourth quadrant group” does the same thing with the last 20 teams in the field.
“Often it requires a bunch of work—who won their regular season, who won their conference tournament, how to sort out the various teams that are both AQs and at the bottom of the at-large field,” Shaheen said. “That’s really a dedicated effort as well.”
After each team is awarded a seed, the rank of those teams determines where to place them, and the committee tries to use an S curve. For instance, the fourth-best No. 2 seed would be placed in the same region with the top overall seed. However, the committee must follow bracketing principles.
For the first time in the process, members look at teams as more than simply as Team X or Team Y. The committee tries to limit travel and avoid matching up teams within the same conference in early rounds.
For a first-timer, this could feel like navigating a maze. The members serve five-year terms, and those terms end at different times, so much of the committee is seasoned.
“I would say in general, committee members report that they spend one year out of that time on committee responsibilities,” Shaheen said. “They’re excellent at what they do. I’m sure they’ll come out with something that deserves a lot of discussion and attention Sunday night.”
In other words, Shaheen believes the bracket is in good hands, and he deserves as much credit as anyone that the process is as organized and streamlined as it is now.
“This event is bigger than any of us—any one person, any one team, any one year,” Shaheen said. “It’s an incredibly humbling experience to have been part of it like I was, and now I get to enjoy it as a fan. … I think change and evolution is a really good thing here.”
Fans may object to the ultimate matchups, but how they get there mostly makes sense. Knowing more about that process is a good thing.
For that, we have Shaheen to thank.