Bracketology

Every year, mid March, the basketball stat-loving fans are knee-deep into "bracketology". This is the pseudo-science that tries to guess who will play in March Madness.

I should back up some. The NCAA is a governing body of college sports in the United States. Each college/university is placed in a "division" based on the size of their school and on their level of athletic prowess. In particular, there is Division 1, 2, and 3. Division 1 is where the best athletes tend to go.

March Madness is the NCAA Division 1 Men's Basketball Championships.

The tournament is in mid to late March and has 65 teams. Each college/university in Division 1 belongs to a conference. Each conference has a way of determining a winner. Sometimes they have a tournament at the end of the season. Most major conferences do this, e.g. the ACC, which stands for the Atlantic Coast Conference had a year-end championship that concluded yesterday, March 15 (Duke beat Florida State).

Duke, for example, got an automatic bid into the tournament by winning the ACC championships and represents the ACC.

The Ivy League looks at the team with the best regular season record (I believe, only against other Ivy League teams). This year, it's Cornell.

There are 31 conferences, each with an automatic bid based on how that conference decides to pick a representative. Why automatic? Many of these conferences are very weak and would not otherwise qualify for March Madness if it weren't for the automatic bid. Fans of basketball often enjoy these smaller (called "mid-major") conferences because the upset, while rare, is entertaining.

There are 34 at-large bids. A committee decides which 34 teams that didn't get an automatic bid qualify. The committee decides, based on a body of work, which teams get an at-large bid. It is the source of nervous anticipation when teams on the "bubble", that is on the verge of making it in or not, wait to find if the committee picked them or not. St. Mary's didn't get in. Arizona did. These two were on the bubble.

In the weeks leading up to Selection Sunday (which was March 15 this year), experts called "bracketologists" (a made up name to sound a bit scientific) try to guess which teams will make the 34 at-large bids.

Now if you add the two numbers: 31 automatic and 34 at-large, you get 65. This is not a power of two. A single elimination tournament should add to 64.

65 came about because a new conference was added to go from 30 to 31. No one wanted the number of at-large bids to go down by 1 especially since the new conference was likely to be very weak.

Usually, March Madness is played over three long weekends. By long weekend, I mean Thursday, Friday, Saturday, Sunday. If a given team is not eliminated, they play two games during this long weekend. 6 wins gives you a national title.

How to deal with 65 then? Two teams play in a play-in game and that is played on the Tuesday (2 days before the first "official" day). The winner usually has the honor of playing a number 1 seed.

Ah, so seeding. Once we get to 64, each team is placed in one of four groups of 16. They are called East, West, Midwest, and South. Within each group, the teams are ranked (or seeded) 1 to 16. They play each other so that 1 plays 16, 2 plays 15, 3 plays 14, and so forth. In other words, the highest seed always plays the lowest seed at each possible round, making their chances better (though not guaranteed) that they will make it to the finals.

Once each of the groups has eliminated down to 1, there are 4 teams remaining, and this is called the "Final Four". In the final weekend, they go from 4 to 2, then 2 to 1 to determine a champion.

For many universities, a good sign of success is making it past the first long weekend. If they win 2 games, they are in the final 16, which is called the "Sweet 16". A measure of how good a basketball program is is how many times they have made the Sweet 16 ever and in the last few years.

Bracketology is interesting because people spend lots of time trying to guess who will make it, and for the most part, they are correct, plus or minus 5-6 picks. The exact seeding is a bit challenging as other factors are put into play.

In particular, high seeds are often placed at a tournament site (picked long before March Madness) near their university so their fans can come and support them. This may seem unfair (it's common wisdom that the more fans that come and cheer, the better a team does), but in the past, lower seeded teams have sometimes been placed closed to home, while the higher seeded opponent is far away and fewer fans attend, thus giving a "home court advantage" to the lower seeded team.

The committee created something called "pods" a few years ago, so they could mix and match teams from different regions (the East, West, Midwest, South from earlier) at the same tournament site so they could help teams stay close to home. This is a tough job, and it affects seeding too.

It's amazing but people spend the last weeks of February and the first weeks of March trying to predict which teams will make it. Time and fastidious attention are critical, and yet it is merely a game.

Oh yeah, betting. This is perhaps the hugest betting that occurs on any event in the US. Most of it is done in small office pools where each person puts in a small amount, between 5 and 20 dollars. It keeps interest up especially in upsets and tiny teams and people who play but know nothing of the sport and do well nonetheless because those that know better pick more upsets than they should.

So it begins.