- Joined
- May 8, 2002
- Posts
- 366,138
- Reaction score
- 43
There are a lot of numbers in front of the men's NCAA Tournament selection committee each year, with the intention that no team be defined by any one of them. There is no threshold to pass for a prospective at-large bid into the field, not even within the NCAA’s own NET ratings.
This, of course, is part of the charm and mystery of March Madness, an event in which the build-up of who’s in, who’s out and who’s seeded where dominates the college basketball conversation. That discussion can’t happen these days without citing at least one of the seven metrics that will appear on the team sheets used by the selection committee this year.
They are broken down into two categories: Predictive metrics and results-based metrics. But the growing role these ratings or rankings take on ahead of Selection Sunday – and the growing number in use by the NCAA (it officially added two more metrics to team sheets for this season) – can sometimes lead to confusion or misconceptions about what aspects of the game each metric is measuring, what data is being used to compare these teams and what it all actually means for the selection and seeding of the NCAA Tournament.
Here’s a breakdown of all seven metrics used by the NCAA tournament selection committee in 2025:
“The NET rankings measure team performance based on efficiency, opponent quality and game location,” according to the NCAA, which considers it the primary sorting tool for evaluating teams ahead of Selection Sunday.
The NET is initially released each season in December, but is designed to be its most accurate in March. It does not use any data from prior seasons or preseason rankings, nor does it take into account scoring margin. This is different from other predictive models. NET also weighs every game equally regardless of date.
The NCAA also utilizes a quadrant system based on its NET rankings to determine the quality of a team’s wins and losses. A team's record in each quadrant is included on its team sheet.
Ken Pomeroy began publishing his ratings at kenpom.com in 2002, and they are viewed as an innovation that created the modern analytics movement in college basketball.
Pomeroy was the first to implement and popularize offensive and defensive efficiency, points per possession and tempo-free stats into his metric. It "added another dimension of information that brought these teams to life a little bit more," he said.
Pomeroy's ratings do not cap margin of victory and incorporate preseason data that is progressively phased out by the end of February and beginning of March as more results from the current season accumulate. The core of the system is the pythagorean calculation for expected winning percentage, a formula originally created by Bill James for use in baseball.
"The purpose of this system is to show how strong a team would be if it played tonight, independent of injuries or emotional factors," Pomeroy wrote on his website. "Since nobody can see every team play all (or even most) of their games, this system is designed to give you a snapshot of a team’s current level of play."
"It likes a team that loses a lot of close games against strong opposition more than one that wins a lot of close games against weak opposition," he added.
ESPN defines its college basketball version of the BPI as "a measure of team strength that is meant to be the best predictor of performance going forward. BPI represents how many points above or below average a team is.” When it debuted in 2011, BPI incorporated aspects of both predictive and results-driven metrics. Today, it is "purely predictive," according to ESPN director of analytics Matt Morris.
BPI uses preseason ratings that hold progressively less weight until it becomes "almost nonexistant," according to the NCAA, and takes into account "points, possessions, opponent strength, game site, distance each team had to travel, team rest and whether the game is played at a high altitude."
"The BPI back in the day is quite a bit different than the BPI now," Morris said. "The best teams are at the top. The worst teams are at the bottom. Straight up, who’s going to win on a neutral court."
Bart Torvik's ratings, created in 2014 and published at barttorvik.com, are appearing on NCAA Tournament selection committee team sheets for the first time this year. His rankings emphasize offensive and defensive efficiency like other predictive models but with an added game script component.
Torvik "omits data after the game mathematically becomes decided," according to the NCAA. "All games played in the previous 40 days count 100%, then degrade 1% per day until they're 80 days old, after which games count 60%."
“It’s adjusted offensive and defensive efficiency with a little bit of this game control aspect added," Torvik said. "It has this unique aspect that some ratings are different if they’ve blown a lot of big leads or come back a lot. That’s where you get different ratings, but it is very similar to KenPom. It’s ultimately just based on who you played and how much you beat them by or lost by.”
"The Kevin Pauga Index metric ranks team resumes by assigning a value to each game played. The best win possible is worth about +1.0, the worst loss about -1.0, and a virtual tie at 0.0," according to the NCAA. "Adjustments are made to each game's value based on location of the game, opponent quality and percentage of total points scored. Game values are added together and divided by games played to determine a team's KPI ranking."
This allows a team's schedule to be sorted by quality of wins and losses. The date of each game does not factor into the KPI rankings.
“KPI was rooted in the why of building a non conference schedule," Pauga said. "How do we quantify at a more precise level as to what the actual impact is of playing this team instead of that team."
ESPN officially defines its Strength of Record metric as "a measure of team accomplishment based on how difficult a team's W-L record is to achieve." The probability of winning each game is based off a team's current BPI rating.
"It packs a punch and tells you so much," Morris said. "It's the chance of an average top-25 team achieving your record or better given your schedule. It’s kind of a mouthful, but basically it’s just how impressive is your record given your schedule.”
In the same vein as SOR, wins above bubble "calculates the expected winning percentage for an average bubble team in each game of a team's schedule," according to the NCAA, which is incorporating the statistics on selection committee team sheets for the first time this year.
WAB breaks down into the amount of wins you have minus the amount of wins an average bubble team would expect to have versus your schedule. That bubble team has been set as the No. 45 team in the current NET ratings based on past data analysis. Seth Burn, a professional gambler, is viewed as the first to conceive of WAB as a metric for rating college basketball teams in 2015, with Torvik popularizing the metric in recent years by incorporating it into his ratings.
"It’s kind of taking win-loss and adjusting for schedule," Alok Pattani, a data science developer for Google who helped create the NET and the NCAA's version of WAB, told NCAA.com. "We talk a lot about strength of schedule, which is important. This is how did you do against that schedule. ... I think wins (above) bubble is a really good advancement to work around some of these issues at the edges of the quad system."
This article originally appeared on USA TODAY: March Madness bracket: What are metrics used for 2025 NCAA tournament?
Continue reading...
This, of course, is part of the charm and mystery of March Madness, an event in which the build-up of who’s in, who’s out and who’s seeded where dominates the college basketball conversation. That discussion can’t happen these days without citing at least one of the seven metrics that will appear on the team sheets used by the selection committee this year.
They are broken down into two categories: Predictive metrics and results-based metrics. But the growing role these ratings or rankings take on ahead of Selection Sunday – and the growing number in use by the NCAA (it officially added two more metrics to team sheets for this season) – can sometimes lead to confusion or misconceptions about what aspects of the game each metric is measuring, what data is being used to compare these teams and what it all actually means for the selection and seeding of the NCAA Tournament.
Here’s a breakdown of all seven metrics used by the NCAA tournament selection committee in 2025:
March Madness 2025: NCAA Tournament predictive metrics
NCAA Evaluation Tool (NET)
“The NET rankings measure team performance based on efficiency, opponent quality and game location,” according to the NCAA, which considers it the primary sorting tool for evaluating teams ahead of Selection Sunday.
The NET is initially released each season in December, but is designed to be its most accurate in March. It does not use any data from prior seasons or preseason rankings, nor does it take into account scoring margin. This is different from other predictive models. NET also weighs every game equally regardless of date.
The NCAA also utilizes a quadrant system based on its NET rankings to determine the quality of a team’s wins and losses. A team's record in each quadrant is included on its team sheet.
- 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 (136-240).
- Quadrant 4: Home (161-353), Neutral (201-353), Away (241-353).
KenPom
Ken Pomeroy began publishing his ratings at kenpom.com in 2002, and they are viewed as an innovation that created the modern analytics movement in college basketball.
Pomeroy was the first to implement and popularize offensive and defensive efficiency, points per possession and tempo-free stats into his metric. It "added another dimension of information that brought these teams to life a little bit more," he said.
Pomeroy's ratings do not cap margin of victory and incorporate preseason data that is progressively phased out by the end of February and beginning of March as more results from the current season accumulate. The core of the system is the pythagorean calculation for expected winning percentage, a formula originally created by Bill James for use in baseball.
"The purpose of this system is to show how strong a team would be if it played tonight, independent of injuries or emotional factors," Pomeroy wrote on his website. "Since nobody can see every team play all (or even most) of their games, this system is designed to give you a snapshot of a team’s current level of play."
"It likes a team that loses a lot of close games against strong opposition more than one that wins a lot of close games against weak opposition," he added.
Basketball Power Index (BPI)
ESPN defines its college basketball version of the BPI as "a measure of team strength that is meant to be the best predictor of performance going forward. BPI represents how many points above or below average a team is.” When it debuted in 2011, BPI incorporated aspects of both predictive and results-driven metrics. Today, it is "purely predictive," according to ESPN director of analytics Matt Morris.
BPI uses preseason ratings that hold progressively less weight until it becomes "almost nonexistant," according to the NCAA, and takes into account "points, possessions, opponent strength, game site, distance each team had to travel, team rest and whether the game is played at a high altitude."
"The BPI back in the day is quite a bit different than the BPI now," Morris said. "The best teams are at the top. The worst teams are at the bottom. Straight up, who’s going to win on a neutral court."
Torvik
Bart Torvik's ratings, created in 2014 and published at barttorvik.com, are appearing on NCAA Tournament selection committee team sheets for the first time this year. His rankings emphasize offensive and defensive efficiency like other predictive models but with an added game script component.
Torvik "omits data after the game mathematically becomes decided," according to the NCAA. "All games played in the previous 40 days count 100%, then degrade 1% per day until they're 80 days old, after which games count 60%."
“It’s adjusted offensive and defensive efficiency with a little bit of this game control aspect added," Torvik said. "It has this unique aspect that some ratings are different if they’ve blown a lot of big leads or come back a lot. That’s where you get different ratings, but it is very similar to KenPom. It’s ultimately just based on who you played and how much you beat them by or lost by.”
Men's NCAA Tournament 2025: March Madness results-based metrics
Kevin Pauga Index (KPI)
"The Kevin Pauga Index metric ranks team resumes by assigning a value to each game played. The best win possible is worth about +1.0, the worst loss about -1.0, and a virtual tie at 0.0," according to the NCAA. "Adjustments are made to each game's value based on location of the game, opponent quality and percentage of total points scored. Game values are added together and divided by games played to determine a team's KPI ranking."
This allows a team's schedule to be sorted by quality of wins and losses. The date of each game does not factor into the KPI rankings.
“KPI was rooted in the why of building a non conference schedule," Pauga said. "How do we quantify at a more precise level as to what the actual impact is of playing this team instead of that team."
ESPN's Strength of Record (SOR)
ESPN officially defines its Strength of Record metric as "a measure of team accomplishment based on how difficult a team's W-L record is to achieve." The probability of winning each game is based off a team's current BPI rating.
"It packs a punch and tells you so much," Morris said. "It's the chance of an average top-25 team achieving your record or better given your schedule. It’s kind of a mouthful, but basically it’s just how impressive is your record given your schedule.”
Wins Above Bubble (WAB)
In the same vein as SOR, wins above bubble "calculates the expected winning percentage for an average bubble team in each game of a team's schedule," according to the NCAA, which is incorporating the statistics on selection committee team sheets for the first time this year.
WAB breaks down into the amount of wins you have minus the amount of wins an average bubble team would expect to have versus your schedule. That bubble team has been set as the No. 45 team in the current NET ratings based on past data analysis. Seth Burn, a professional gambler, is viewed as the first to conceive of WAB as a metric for rating college basketball teams in 2015, with Torvik popularizing the metric in recent years by incorporating it into his ratings.
"It’s kind of taking win-loss and adjusting for schedule," Alok Pattani, a data science developer for Google who helped create the NET and the NCAA's version of WAB, told NCAA.com. "We talk a lot about strength of schedule, which is important. This is how did you do against that schedule. ... I think wins (above) bubble is a really good advancement to work around some of these issues at the edges of the quad system."
This article originally appeared on USA TODAY: March Madness bracket: What are metrics used for 2025 NCAA tournament?
Continue reading...
