How the NET Ranking Works (And What It Gets Right and Wrong)

Every March, a team's NCAA Tournament hopes get summarized in a phrase: "they're 6-4 in Quadrant 1." That shorthand comes from the NET — the NCAA Evaluation Tool — the ranking the selection committee uses to sort men's college basketball teams. It's official, it's influential, and it's widely misread as a "power ranking." It isn't quite that. Here's what the NET is, how the quadrants work, and what it gets right and wrong.

What the NET replaced, and why

From 1981 through 2018 the NCAA leaned on the RPI, a crude formula that was three-quarters opponents' record. It was easy to game: schedule the right cupcakes and middling teams could inflate their profile. In 2018-19 the NCAA replaced it with the NET, built with help from Google Cloud and reworked since. The goal was a more honest sorting tool — one that values who you beat and how, not just how many.

What goes into the NET

The modern NET blends two ideas:

• Team Value Index — a results-based measure that credits each game by the quality of the opponent and the location (a road win counts more than the same win at home).
• Adjusted net efficiency — points scored minus allowed per possession, adjusted for opponent strength and location, with scoring margin capped (around 10 points) so blowing out a bad team by 40 doesn't help more than winning by 12.

Crucially, the NET does not use a team's win-loss record directly, and it doesn't care about the order of games — a November loss and a February loss are weighted the same. It's a season-long body-of-work number, not a momentum tracker.

The quadrant system

The committee rarely talks about a team's NET number itself. It talks about quadrant records, because where the NET matters most is in classifying every game by opponent quality and site. A win over the same team is worth more on the road than at home — so the bands shift by location:

This table is the secret decoder ring for bracketology. A "Quad 1 win" — beating a top-30 team at home, a top-50 team on a neutral court, or a top-75 team on the road — is the currency of an at-large bid. A "bad loss" almost always means a Quad 3 or Quad 4 defeat. When you hear a bubble team is "1-7 in Quad 1, 0 bad losses," you now know precisely what that means.

What the NET gets right

• It rewards scheduling up. Because road and neutral wins over good teams are worth most, the NET nudges teams to play real non-conference games instead of buy games. That's good for the sport.
• It caps margin. Running up the score doesn't help, which removes an ugly incentive the old systems created.
• It's location-aware and order-blind. A quality win is a quality win whether it happened in November or February.

What it gets wrong (or gets blamed for)

• People treat it as a seed list. It isn't. The NET sorts team sheets; humans on the committee still decide. A team's NET rank and its eventual seed often differ by several lines.
• Quadrant cliffs are arbitrary. An opponent ranked 31st versus 30th can flip a marquee win from Quad 1 to Quad 2 — a discontinuity that can swing a résumé for reasons outside your control, since your opponent's NET keeps moving all season.
• It blends predictive and results-based ideas. Adjusted efficiency is predictive (how good are you?), while the Team Value Index is results-based (what did you do?). Mixing them means the NET is neither a pure power rating nor a pure résumé — which is why the committee looks at it alongside other metrics, not instead of them.
• Mid-majors can get squeezed. A great team in a weak league struggles to find Quad 1 chances at all (more on that in our mid-major at-large math piece).

How to read a team sheet like the committee

Skip the single NET number and read the quadrant rows. Strong at-large profiles look like: several Quad 1 wins, a winning or near-even Quad 1-2 record, and no Quad 3-4 losses. Weak ones pile up Quad 2-3 wins that look fine in the standings but carry little weight. The NET isn't a verdict on who's best — it's a consistent way to compare what teams actually did. Used that way, it's the most useful three letters in March.