Grab a hot chocolate and pull up a chair by the fire. We’re explaining TRACR, expected quad wins and record value – three staples of our college basketball analysis.
Let’s start with the foundation of all this, which is our TRACR model.
TRACR (Team Rating Adjusted for Conference and Roster) is a net efficiency metric trained on over 15 years of college basketball data that looks at how good a team is on every possession relative to how good its opponent is. And the accumulation of that gives us an offensive rating and a defensive rating – the difference between those gives us a team rating.
Factoring these inputs together produces an accurate measurement into how men’s and women’s teams have performed and will perform during the season.
For example, a team with a TRACR of 0.0 is considered an average team and a team with a rating in the positive indicates how much it would outscore the average team. So let’s say that San Diego State has a TRACR of 0.0 and Arizona has a rating of 14.0. If San Diego State and Arizona played each other at a neutral site, the most likely outcome would be a Wildcats victory by about 14 points.
Our preseason TRACR rankings factor in things like the previous year’s performance, recruitment classes and ratings, returning minutes to the team and the caliber of transfer players added to the program.
The model removes some of the bias that AP Top 25 voters have when it comes to name recognition of the blue-blood programs and overvaluing teams that went far in the previous year’s NCAA Tournament.
Our model can also help us get a clear idea of which conferences as a whole are the best in the country by average TRACR, which is updated throughout the season. As an example, the Big 12 was the No. 1 league during the 2023-24 season with an average team rating of plus-25.7.
2023-24 TRACR Conference Rankings (as of Feb. 2)
- Big 12 – 25.7
- SEC – 20.53
- Big Ten – 20.45
- Big East – 18.9
- ACC – 18.3
- Pac-12 – 15.1
- Mountain West – 14.2
Expected Quad Wins
In recent seasons, the NCAA Tournament Selection Committee has been leaning on “quad wins” quite a bit.
That’s because the NCAA developed a computer model ahead of the 2018-19 season called the NET (NCAA Evaluation Tool) that calculates team ratings used to rank every D1 basketball team and replaced the RPI in the committee’s toolbox.
Using the NET rankings, team wins and losses have been sorted into four quadrants. Those are organized based on game location and the opponent’s NET ranking, helping to put a greater value on victories on neutral courts and in true road environments.
Here’s how that looks:
- Quad 1: Home games vs. 1-30, neutral games vs. 1-50 teams, road games vs. 1-75
- Quad 2: Home vs. 31-75, neutral vs. 51-100, road vs. 76-135
- Quad 3: Home vs. 76-160, neutral vs. 101-200, road vs. 136-240
- Quad 4: Home vs. 161-351, neutral vs. 201-351, road vs. 241-351
Simply put, Quad 1 wins are important because they represent the best competition the nation has to offer. And Quad 4 victories, for example, have less value because they come against opponents a good team should beat.
Quad records are always based on current NET ratings, meaning that if a team falls out of a quadrant threshold, it will impact the records of every team that played them.
That complicates matters for bracketologists and college basketball experts attempting to project the field of 68. But using TRACR, we’ve developed a metric called expected quad wins that projects a team’s record in Quads 1-4 over the rest of the season.
Since TRACR already predicts overall record, it can also reveal how the team will fare in those respective quadrants – one of those major factors in determining who should get in the NCAA Tournament and who should be left out.
We’ve also developed something called record value, which estimates how likely an average TRACR top-25 team would match or exceed a team’s expected record at the end of the regular season. This process can give us an indication of how strong a team’s record is: The lower the team’s record value, the rarer that record is.
For example, let’s say Wisconsin’s expected record at the end of the regular season is 22.9-8.1. For simplicity’s sake, let’s say they finish 23-8 on average. So how many times would the average TRACR top-25 team play through that same schedule and go 23-8 or better? Wisconsin’s record value is 61.4, meaning that a top-25 team would match or exceed that record about 61.4% of the time.
Each teams’ rankings, projected record both overall and in conference play, expected wins for each quad and record value are updated throughout the season in our college rankings and predictions page.