Bracketology, Explained: How the Committee Picks and Seeds the 68

Every March, a small group of people lock themselves in a room, stare at hundreds of “team sheets,” and produce the one document that defines the college basketball postseason: the bracket. The rest of us spend weeks guessing what they’ll do — an entire cottage industry called bracketology. This piece explains the real machinery behind it: how the men’s NCAA Tournament field is built, what the selection committee actually weighs, how teams are seeded on an “S-curve,” and the bracketing rules that move teams around once the seeds are set.

The one-sentence version: 32 teams earn their way in by winning a conference; 36 more are chosen by a committee reading résumés; and then all 68 are ranked, seeded, and slotted into a bracket under a thick rulebook of geography and fairness constraints.

The structure of the field: 68 teams, two paths in

The men’s tournament field is 68 teams, and they arrive by two completely different routes:

• 32 automatic bids. Each Division I conference sends its champion — in nearly every league, the winner of the conference tournament. Win your league’s tournament and you are in, full stop, regardless of your overall record. This is the path that makes Championship Week so wild.
• 36 at-large bids. The remaining spots are handed out by the selection committee, which picks the best teams that did not win an automatic bid. These are the “at-large” teams, and the last few in are the famous “bubble.”

Add them up — 32 + 36 = 68 — and you have the field. Because the bracket plays down from 64, the eight lowest-seeded teams (a mix of the weakest automatic qualifiers and the last at-large teams in) meet in the First Four play-in games before the round of 64.

What the committee weighs

The committee does not rank teams by a single formula. It reads a team sheet for each at-large candidate — a standardized one-page profile — and weighs a blend of results-based and predictive information. The centerpiece is the NET ranking (the NCAA Evaluation Tool) and, more than the NET number itself, the quadrant records it generates. We cover that machinery in detail in our NET ranking explainer; the short version is that every game is bucketed into Quadrant 1 through 4 by opponent quality and location, so a road win over a top-75 team counts far more than a home win over a weak one.

Around that core, the committee considers:

• Quadrant 1 and 2 wins — the currency of an at-large bid. Strong profiles stack good wins; weak ones pile up wins that look fine in the standings but carry little weight.
• Bad losses — Quadrant 3 and 4 defeats are the fastest way to fall off the bubble.
• A mix of metrics — both results-based rankings (what you did) and predictive rankings (how good you appear to be). The committee deliberately looks at several, not one.
• Road and neutral-court wins — winning away from home is hard and is credited accordingly.
• The “eye test” and context — injuries, who was available for which games, and how a team is actually playing. Human judgment is the final layer.

This is also where the math gets brutal for great teams in weak leagues: if your conference offers almost no Quad 1 chances, it is hard to build the kind of résumé the committee rewards — the squeeze we break down in the mid-major at-large math.

Seeding the field: the S-curve

Once the 68 teams are chosen, the committee does something distinct from selecting: it ranks the entire field 1 through 68, a list known as the S-curve (or “seed list”). Seeds then fall out of that ranking in groups of four. Conceptually:

The S-curve is also used to balance the four regions. The overall No. 1 team is placed in the strongest position, then the regions are filled by “snaking” the list so that, in theory, each region is roughly equal in total strength — the top overall 1-seed shouldn’t share a region with the strongest 2-seed, and so on. The goal is four balanced quadrants, not four random ones.

Bracketing principles: geography and fairness

Seeding tells you which line a team is on; bracketing decides where it actually goes. The committee follows a rulebook of principles, sometimes bending the pure S-curve to satisfy them. The big ones:

• Place teams near home. Higher seeds are slotted, when possible, into the region and early-round sites closest to their campus — a reward for a strong season.
• Avoid early conference rematches. Teams from the same conference are kept apart in the early rounds (the constraint loosens the deeper you go), so the bracket doesn’t reproduce regular-season matchups in the first weekend.
• Protect the top seeds. The bracket is built so the best teams have the most favorable feasible path; if a team must be moved off its “true” seed line for other rules, the committee tries to swap it with the team nearest it on the S-curve to keep the field fair.

These principles routinely conflict — keeping a team near home might force a conference rematch, for instance — and resolving those conflicts is most of the committee’s bracketing work.

Bid stealing: why the bubble holds its breath

Here is the dynamic that makes Championship Week unbearable for bubble teams. There are exactly 36 at-large spots. Every one of those spots assumes that the automatic bids go to teams that would have made the field anyway. When a non-favorite wins its conference tournament — a team that had no realistic at-large case — it claims an automatic bid that, in effect, removes one at-large spot from the pool. That’s a “bid steal” (or “bid thief”).

The arithmetic is unforgiving: each bid steal shrinks the at-large pool from 36 toward 35, 34, and so on, bumping the last teams off the bubble. This is why fans of bubble teams root, with genuine anxiety, for the favorites to win conference tournaments in leagues they have no other stake in. One upset two time zones away can end their season.

How to read a bracket projection like the committee

When you see a projected field, ignore the bottom-line seed at first and read the inputs: quadrant records, the count of Quad 1 wins, whether there are any bad losses, and the road/neutral profile. A team with several Quad 1 wins and no bad losses is safe even if a metric or two is lukewarm; a team with a glossy record built on Quad 3 wins is fragile. Then remember the live variable nobody controls — bid stealing — and you’ll understand why the bracket can look settled on Saturday morning and chaotic by Sunday. The committee isn’t guessing; it’s applying a consistent process to a moving target. Bracketology is just the art of reading that process before it finishes.