Decoding the Metrics: A Guide to the NCAA Tournament Selection Committee’s Tools

The annual spectacle that is March Madness is fueled by anticipation, bracketology, and passionate debate. At the heart of this frenzied period lies the NCAA Tournament selection committee, tasked with the unenviable job of distilling hundreds of teams down to a field of 68. While many fans rely on gut feelings and casual observations, the committee employs a more data-driven approach, utilizing a variety of metrics to evaluate teams and ultimately determine which squads deserve a coveted spot in the tournament and where they should be seeded.

While the allure of March Madness lies in its unpredictability, the selection process itself is becoming increasingly transparent, with the NCAA sharing the tools it uses to make its decisions. This article provides a breakdown of the seven key metrics the committee will utilize in 2025, shedding light on what they measure, how they work, and what they mean for your favorite team’s chances.

The Importance of Multiple Perspectives

It’s crucial to understand that no single metric dictates a team’s destiny. The selection committee emphasizes a holistic approach, considering a range of factors and avoiding reliance on any single number. There isn’t a magical threshold a team must cross to secure an at-large bid, even within the NCAA’s own NET rankings. The committee aims to avoid teams being defined solely by a single statistic.

These metrics are broadly categorized into two types: predictive and results-based. Predictive metrics attempt to forecast future performance, while results-based metrics focus on what a team has already accomplished. The increasing reliance on these ratings and rankings, coupled with the NCAA’s ongoing expansion of the metrics it uses, can lead to confusion and misinterpretations. So, let’s delve into the specifics:

The Seven Metrics of March Madness

1. NET (NCAA Evaluation Tool)

The NET rankings serve as the primary sorting tool for the committee. It measures team performance based on three core components: efficiency, opponent quality, and game location. The NET is updated throughout the season, with its most accurate representation occurring in March.

Unlike some predictive models, the NET excludes data from previous seasons or preseason rankings and does not factor in scoring margin. Every game is weighed equally, regardless of when it was played. Furthermore, the NET also underpins the quadrant system, which categorizes wins and losses based on the opponent’s NET ranking and the game location. A team’s record in each quadrant provides valuable context about the quality of its wins and losses.

2. KenPom (Ken Pomeroy’s Ratings)

Ken Pomeroy’s ratings, found at kenpom.com, are considered a cornerstone of modern college basketball analytics. Pomeroy pioneered the use of offensive and defensive efficiency, points per possession, and tempo-free statistics.

Unlike the NET, Pomeroy’s ratings do not cap margin of victory and incorporate preseason data that gradually diminishes as the season progresses. The system uses a pythagorean calculation to estimate a team’s expected winning percentage, aiming to provide a snapshot of a team’s current level of play, irrespective of injuries or emotional factors. This system favors teams that lose close games against strong opponents over teams that win close games against weak opponents.

3. BPI (Basketball Power Index)

ESPN’s BPI is a predictive metric designed to be the best indicator of a team’s future performance. It represents how many points above or below average a team is. While initially incorporating both predictive and results-driven elements, the BPI is now purely predictive.

BPI incorporates preseason ratings that gradually lose weight as the season goes on. It also takes into account factors like points, possessions, opponent strength, game site, distance traveled, team rest, and altitude. The BPI aims to predict the outcome of a neutral court matchup between any two teams.

4. Torvik (Bart Torvik’s Ratings)

Bart Torvik’s ratings, available at barttorvik.com, are a more recent addition to the committee’s arsenal. Like other predictive models, Torvik emphasizes offensive and defensive efficiency, but with an added element of game script.

Torvik’s ratings exclude data from games after they are mathematically decided. Games played within the past 40 days count 100%, gradually decreasing by 1% per day until 80 days old, after which they count 60%. This unique approach can differentiate teams that consistently blow leads or stage comebacks, even if their overall statistics are similar to other teams.

5. KPI (Kevin Pauga Index)

The Kevin Pauga Index (KPI) ranks team resumes by assigning a value to each game played. The best possible win is worth approximately +1.0, the worst loss about -1.0, and a virtual tie at 0.0. These values are adjusted based on game location, opponent quality, and the percentage of total points scored.

A team’s KPI ranking is calculated by adding together the values of all its games and dividing by the total number of games played. This allows for a ranking of teams based on the quality of their wins and losses, but does not factor in the date of each game. KPI seeks to quantify the impact of playing one team versus another when building a non-conference schedule.

6. SOR (Strength of Record)

ESPN’s Strength of Record (SOR) is a measure of team accomplishment based on the difficulty of a team’s record. The probability of winning each game is based on the team’s current BPI rating.

SOR assesses how impressive a team’s record is given its schedule. It essentially measures the chance of an average top-25 team achieving the same record or better, given the same schedule.

7. WAB (Wins Above Bubble)

Wins Above Bubble (WAB) calculates the expected winning percentage for an average bubble team in each game of a team’s schedule. This metric is calculated by subtracting the number of wins an average bubble team would expect to have against a given schedule from the team’s actual win total.

The "bubble team" is defined as the No. 45 team in the current NET rankings, based on historical data. WAB effectively adjusts a team’s win-loss record based on the strength of its schedule. It offers a way to address potential issues at the edges of the quadrant system by providing context about how a team performed against its schedule.

The Human Element Remains

While these metrics provide valuable insights, it’s important to remember that the selection committee is comprised of human beings who also consider factors not easily quantified, such as injuries, team chemistry, and recent performance trends. The committee watches games, gathers information, and engages in robust discussions before making its final decisions.

The selection process is a complex balancing act, weighing both objective data and subjective observations. Understanding these metrics can enhance your appreciation of the process and help you engage in more informed discussions about which teams deserve a spot in the Big Dance. Ultimately, though, the magic of March Madness lies in its inherent unpredictability, reminding us that even the most sophisticated algorithms can’t fully capture the drama and excitement of college basketball.