What I learned from the NCAA’s mock tournament selection

Bracketologist Russell Steinberg visited Indianapolis to take part in a mock tournament selection exercise. Here’s what the committee taught him.

INDIANAPOLIS — The annual NCAA women’s basketball mock tournament selection exercise is an opportunity for coaches to get an inside look at how the selection committee comes up with its bracket. As one of a handful of media members present, I had the opportunity to do the same.

More from Bracketology

• Women’s basketball bracketology: Bid-stealers knock teams off the bubble
• Women’s basketball bracketology: Selection Committee reveals UCLA is a 2 seed
• Women’s basketball bracketology: Yes, Northwestern is a 2 seed
• Women’s Basketball Bracketology: The case for Maryland as a 1 seed
• Women’s Basketball Bracketology: NC State ascends to the 1 line

The basic outline of the process is publicly available, and I used it to the best of my ability last year while creating my mock brackets for High Post Hoops. It’s a handy start to an aspiring bracketologist or anyone who wants to better understand why teams are placed where they are. But there are some elements that cannot be covered in a handbook quite like they can be demonstrated in person.

That’s why I went to Indianapolis. If my job is to mirror the committee as best I can throughout the season, this is one way to guarantee I can do it better.

Here are a few things I learned:

RPI matters a ton, for now

There is a lot to criticize about the RPI, and the committee recognizes that. In fact, vice president of women’s basketball Lynn Holzman said during a pre-mock-selection presentation that the sport is looking into developing an alternative metric that may better measure the quality of teams. The men’s game is doing that right now and should have something to present in the coming weeks. The women’s game will follow that — not in time for the 2019 NCAA Tournament, but within the next few years. Holzman says that while it may include some elements of the men’s formula, it will not be the exact same one because the games aren’t starting from the same place.

So for now, fair or not, look at the RPI. The selection software that the committee uses throughout the process sets RPI as a starting point for rankings and as the reference for records against top 50 teams, top 100, etc. In fact, RPI and strength of schedule were the two data points we used in our selection more than any other, and it wasn’t close.

Injuries matter too, but only once the field is set

This was said a number of times throughout the exercise. The actual team selection is made based on what a team has done in the regular season — so that’s win/loss record, strength of schedule, good wins, bad losses, etc. Seeding is based on what the team is right now. So if a player got hurt in the conference tournament and that may have a major impact on their NCAA Tournament performance, it will be reflected in the team’s seed. It will not, however, be reflected in whether the team makes the field.

Everything is voted on, everything is discussed

This was, perhaps, the most surprising takeaway. Sure, it’s outlined in the handbook, but even so, it’s hard to understand exactly how much each detail is debated before it becomes a part of the bracket. Try to keep up here:

• First, all committee members submit an initial ballot of teams that they believe should definitely be in the field as at-larges and teams that they believe should be considered.
• If a team appears on all but one initial ballot as an at-large, it is automatically put into the field. All other teams are put on the board to consider.
• Next, each committee member ranks its top eight remaining teams.
• The four teams receiving the most votes are placed into the field. The rest go back to the board for consideration.
• This process continues until all 32 at-large teams have been picked.

If, at any point, a committee member takes issue with a team that has been selected, he or she can stop the process and ask for a discussion and potentially another vote. If that team is removed from the field, it goes back to the board for consideration to be voted on again later. This can happen at any point about any detail — selection, seeding, location, you name it. This ensures that EVERY eligible team (any team finishing .500 or above that is not already banned from the postseason) has a shot, so long as committee members think their resume merits discussion.

The committee does its homework 

In theory, we should know this already — that members watch and track a bunch of games throughout the year. Each committee member is expected to be the resident expert on several primary conferences and have extensive knowledge about several secondary conferences. From the beginning of the year, they take periodic conference calls from coaches in their leagues, track the progress of teams that may make the field, track injuries or other circumstances off the court that may explain particular results, and more. When the committee convenes to select the field, they know so much more than what the data says.

Teams need to schedule the best games possible that they think they could win

This was something that committee chair Rhonda Bennett said, and it probably stuck with me the most. Not all wins are equal, but most wins are still better than most losses. At least they should be (don’t get me started about Oklahoma’s inclusion in the field this year). There’s a certain amount of guesswork that goes into this, of course, since we never know exactly how good a team is going to be until the season starts. But it’s all about giving your team an opportunity to be successful. If, for example, a team in the Atlantic 10 thinks it will have an at-large case, it’s probably not a great idea to schedule eight non-conference games against likely top 20 teams (unless they’re REALLY confident). If that A-10 school wants to get the committee’s attention, maybe it’s smarter to try and get a game against two or three top 20 teams, a couple that might be in the Top 25 conversation, and a few more against mid-majors that should also finish near the top of their leagues. The result would hopefully be a strong win-loss record to go with a solid RPI and strength of schedule. That’s preferable to a whole bunch of losses, a lower RPI as a result, and a great strength of schedule.

The bracket has serious limitations. This may or may not be a problem.

The committee has a series of limitations when it comes to bracketing, and there’s logic behind all of them. Some are hard-and-fast rules, others are merely guidelines that they try to abide by but sometimes have to break.

Example of a rule: Under no circumstances can a team play a regional on a court where they played more than three home games this season. So, if Louisville is hosting a regional, the Cardinals need to go somewhere else.

Example of a consideration: The committee will try to avoid conference matchups before the regional final. However in the mock exercise we did, there were five ACC teams on the top four seed lines, making this impossible. Though the committee will sometimes move a team up or down a seed line to make things work, that would have been of no help here. Making a 4 seed a 5 seed, for example, would have kept that team in the same pod AND would have unnecessarily penalized them by removing them as a host. The committee will also never knock a 1 seed down to a 2 seed. The result: We had a potential all-ACC Sweet 16 game. Not ideal, but sometimes it happens.

Nothing listed above is a major problem. Look to the actual 2018 bracket to see an issue that could actually be debated.

Per the guidelines of the committee, the Spokane regional shaped up exactly as it should have. Notre Dame was the last 1 seed on the S-Curve (or true seed list), so it was the final 1 seed placed in the bracket and went to the only remaining region: Spokane. Oregon, meanwhile, was a 2 seed and, as the rules outline, should have been placed in the regional that made the most geographic sense: also Spokane. The problem arose in the Elite Eight when those two teams met. Notre Dame, as the 1 seed, should have had the advantage, but was forced to play a virtual road game against a Ducks team located less than 500 miles away with a fanbase that permeated its home state and seeps into Washington. Notre Dame won the game anyway, but there would have been no way of knowing that at the time.

Geography matters so much more than I thought

Admittedly, my bracketing process was flawed this season. I took my true seed list and bracketed based only off of that, with the strongest 2 seed paired with the weakest 1, the weakest 2 paired with the strongest 1, etc. Then from there, I adjusted for typical bracket limitations such as avoiding conference matchups. The S-Curve, however, is not the starting point in reality. It is merely the order in which teams are placed. So while the strongest 2 seed will be the fifth team overall to be placed into the bracket, it won’t necessarily go with the weakest 1 seed. It will instead go to the region that makes the most geographic sense, provided the 1 seed in that region is not in the same conference.