With less than two months until Selection Sunday, Duke is projected as a No. 1 seed in the majority of NCAA Tournament brackets. Learn more about what factors into a team's seed line in March:
Overview
The NCAA Tournament — also known as March Madness — is the final battleground of the college basketball season and the pathway to claim the title of national champion. Hundreds of Division I programs compete every year for their spot in the single-elimination bracket and the chance to end their season with a banner.
The tournament begins with four play-in games called the First Four: two between No. 11 seeds, the final four at-large bids, and two between No. 16 seeds. The winner of each game advances to the more traditional 64-team bracket. Each team is assigned a seed and a region. There are four regions — East, Midwest, West and South — and seeds ranging from No. 1 to No. 16. Teams that earn more favorable seeds can expect to play weaker teams in their early-round matchups.
Selecting the field
The critical process of fielding and seeding the 68-team tournament is overseen by a selection committee, one on the men’s side and one on the women’s. Each committee is composed of a mix of athletic directors and conference commissioners from throughout Division I, though these members may not contribute to the discourse surrounding their own schools.
Teams can earn a berth in the tournament through one of two ways:
- Automatic bid: The NCAA consists of 32 Division I conferences, but since the Pac-12 does not get an automatic bid this season, there are only 31 automatic bids. A team that wins its respective conference tournament will punch its ticket to the NCAA Tournament, regardless of record or strength of schedule in the regular season. The lowest seeds (largest numbers) are typically reserved for automatic bids for teams from lower-level conferences.
- At-large bid: Teams that do not receive an automatic bid are “at-large.” The selection committee is responsible for granting an at-large bid to who they view to be the 37 best teams who did not win their conference tournament.
Much of the discussion about the at-large bids surrounds bubble teams. A team is on the bubble if it is uncertain whether it will make or miss the tournament. The bubble is partitioned into four components. Most likely to make the tournament are the “last four byes” followed by the “last four in.” Teams on the bubble that are projected to miss the tournament lie in the “first four out” or “next four out" categories.
Bracketologists including ESPN’s Joe Lunardi and NCAA.com’s Andy Katz provide frequent bracket predictions of the tournament field throughout the season, analyzing what bubble teams they believe will make the field and where teams will be seeded.
Seeding the field
A central concern for the committee is exactly how to rank teams for the sake of granting at-large bids and determining seeding. For the 2024-25 season, the process differs between the men’s and women’s committees. The women’s committee uses only the NET rankings, detailed below, in conjunction with more traditional metrics like overall record, strength of schedule and head-to-head outcomes. The men’s side utilizes these measures in conjunction with six other advanced metrics — a mix of predictive models and resume-based analytics. Although exact justification for a team’s seeding is not provided by the selection committee, expect the results to closely mirror rankings across these platforms.
NET rankings
At the core of resume evaluation are the NCAA Evaluation Tool (NET) rankings. The NET rankings, created in 2018 and consistent across men’s and women’s basketball, combine Team Value Index (TVI) and adjusted net efficiency to yield a consolidated number driven both by results and quality of play. Both aspects of this statistic are weighted based on the playing environment — performance on the road is rewarded — and opponent strength.
Notably, the NET algorithm is agnostic of game date and order. Early-season performance matters just as much as the final games of the regular season.
A key extension of the NET rankings is the quadrant system for categorizing a team’s wins and losses. The quality of a given win or loss can be defined by its quadrant, assigned according to the game’s location and the NET ranking of the opponent. The cutoffs for these quadrants are as follows:
- Quadrant 1: Home 1-30, Neutral 1-50, Away 1-75
- Quadrant 2: Home 31-75, Neutral 51-100, Away 76-135
- Quadrant 3: Home 76-160, Neutral 101-200, Away 135-240
- Quadrant 4: Home 161-353, Neutral 201-353, Away 241-353
It is important to note that the quadrant of a given game can change as the opponent's NET ranking changes. For example, Duke defeated Pittsburgh in Durham Jan. 7. At the time, this was considered a coveted Quad 1 win. However, the Panthers’ recent losing streak yielded a drop to No. 32 in the NET rankings. Now, the Blue Devils’ win is only Quad 2, because it was a home win against a team ranked between 31 and 75.
Duke’s men’s team sits at No. 2 in the NET rankings, trailing only Auburn and boasting impressive Quad 1 wins over Arizona, Auburn, Louisville and SMU. The team is also undefeated in games outside Quad 1. The women’s team is No. 8 in the NET rankings, holding a respectable 3-3 record in Quad 1 opportunities.
Due to a weaker ACC this season, the men’s team currently has only four remaining Quad 1 opportunities on its schedule: at Wake Forest, at Clemson, against Illinois on a neutral floor in Madison Square Garden and at North Carolina. Even by winning all these games, the Blue Devils would fall short of the Tigers’ staggering 10 Quad 1 wins. They also have a plethora of opportunities for more throughout their grueling SEC schedule.
Get The Chronicle straight to your inbox
Sign up for our weekly newsletter. Cancel at any time.
KenPom
The first of the metrics used only in men’s basketball, KenPom is a ratings model centered around the comparison of teams based on offensive and defensive efficiency devised by statistician Ken Pomeroy. The concept is to weigh a team’s past performances based on the strength of the opponent to predict how it would perform against the average Division I team.
The core KenPom analytics are adjusted offensive efficiency, adjusted defensive efficiency and adjusted efficiency margin. Adjusted offensive efficiency is a team’s expected points scored per 100 possessions against an average Division I opponent. Similarly, adjusted defensive efficiency is a team’s expected points allowed per 100 possessions against an average Division I opponent. The difference between these two values gives a team’s adjusted efficiency margin.
Duke is in the midst of a historic season from a KenPom perspective — it holds a 124.6 offensive rating that is fifth in the country while also measuring up at 87.8 on the defensive end, good for second. The adjusted efficiency margin is 124.6 minus 87.8, or 36.8. With this expected 36.8-point edge over a 100-possession competition with an average Division I team, the Blue Devils sit atop the ladder.
KPI
KPI, originating from the analyses of Michigan State University Associate Athletic Director Kevin Pauga, ranks a team’s records by evaluating individual wins and losses on a scale of -1.0 to 1.0. A 1.0-rated game constitutes what the model deems a theoretical “best possible win,” whereas a -1.0 game captures a “worst possible loss.” The average of these scores gives a weighted win percentage used to produce the KPI rankings.
Note that performances outside what the model expected to be possible could result in games rated outside of the -1.0 to 1.0 range. For example, Ole Miss earned a 1.16 KPI rating in its resume-defining 74-64 win at Alabama Jan. 14 — the highest single-game rating this season.
The quality of a given win or loss is based on factors like location, quality of opponent, scoring margin and the pace at which the game was played. The Blue Devils are weaker in KPI relative to other metrics — they hold the eighth spot with their best game a 1.02-rated win at Louisville.
Basketball Power Index
The Basketball Power Index, BPI, is an ESPN-devised model similar to KenPom in that it ranks teams based on an adjusted offensive and defensive efficiency measure relative to an average Division I team. A team’s BPI is the sum of its BPI offense and BPI defense. It indicates the expected point margin per 70 possessions against an average opponent at a neutral site.
Duke’s 24.3 BPI is good for first in the country. This means, per BPI, the team is expected to outscore an average Division I team on a neutral court by 24.3 points for every 70 possessions played.
Strength of Record
More commonly referred to as SOR, this metric is resume-oriented and is intended to gauge the likelihood that a given team’s record would be achieved based on its schedule. SOR is based on a model that simulates the season 10,000 times and estimates the chance a typical 25th-ranked team would have a record at least as good as a given team’s record. This statistic is results-oriented; it typically favors teams with a stronger strength of schedule compared to efficiency-based metrics like KenPom or BPI that reward teams for large winning margins against weak teams. SOR is also unique in that it factors in variables like travel time, rest time and altitude of the play site. The Blue Devils are currently fifth in SOR.
Torvik
New to the selection committee’s evaluation criteria for the 2024-25 season, Torvik refers to the predictive metrics of BartTorvik.com. At the core of Torvik is the T-Rank, a comparative measure of efficiency on both ends of the court that is calculated very similarly to both KenPom and BPI.
T-Rank captures a few notable phenomena to distinguish itself from the other efficiency-based metrics. First, it incorporates a team’s average lead throughout a game and eliminates garbage time — when a game is all but decided — in an attempt to phase out so-called stat padding that would inflate a team’s ranking. T-Rank also employs a recency bias that favors results in contests within the last 40 days while also gradually phasing out preseason expectations once enough games have been played.
According to Torvik, Duke has a 97.96% chance of beating an average Division I team. This puts it at third in the country, behind only Auburn and Houston.
Wins Above Bubble
Also known as WAB, this measure is also new for the 2024-25 season. WAB is another resume-based statistic that compares a team’s performance against its schedule relative to how a bubble team — defined as a team ranked 45th in NET — would be expected to perform against the same schedule. This metric utilizes the NET rankings to gauge the strength of schedule. The Blue Devils sit 4.15 wins above the bubble — sixth in the NCAA.
Note: All statistics and rankings are as of Jan. 22.