After decades of college basketball fans complaining about the use of an outdated and unfair rating percentage index (RPI) in the NCAA tournament selection process, a new team ranking system called the NCAA Evaluation Tool (NET) was implemented for this season.
Thus far, it's unclear whether this is an improvement.
We won't bore you with the details of how the NET is calculated because, well, we couldn't if we wanted to. The NCAA has been tight-lipped about how the sausage is made. When the switch to the NET was originally announced in August, we were told there would be five components:
- Team Value Index (the results-oriented portion of the formula, which factors in strength of schedule)
- Net Efficiency (points scored and points allowed on a per-possession basis)
- Winning Percentage (overall D-I winning percentage)
- Adjusted Win Percentage (accounting for locations of games)
- Scoring Margin (capped at 10 points)
That's all the transparency we have been given, and it left more questions than answers.
There's no indication of how the five factors are weighted, and there is no NET score as far as the public has been shown. We can look at RPI scores or vote totals in the AP poll and get a sense of how wide or narrow the gap is between two teams, but we have no clue with NET. The NCAA simply puts out an updated ranking at some random time between 9 a.m. ET and 4 p.m. each day.
That secrecy might be the most annoying part of it all, and it is already resulting in some unorthodox late-game coaching decisions:
While no one understands the minutiae of the changes, the general idea is that margin of victory matters now—and it matters way too much as far as my two cents are concerned.
Back in the RPI days, a one-point win at the buzzer looked the same as a 40-point blowout. As a result, there were vast differences when comparing RPI rankings to predictive analytics such as KenPom.com, ESPN's BPI and Sagarin, which are grounded in net efficiency and margin of victory.
But it was clear from the first NET rankings release Nov. 26 that too much emphasis was being placed on scoring margin. Ohio State was No. 1, thanks in no small part to home wins over Purdue Fort Wayne, South Carolina State, Samford and Cleveland State by an average margin of 29.8 points.
You might be thinking: "But the NCAA said the scoring margin would be capped at 10 points, so who cares if teams are winning by 10 or 50?"
That's a great point, but there is no cap on net efficiency (as far as we know), which is just scoring margin presented at a per-possession level. And with no knowledge of how each of the five factors is weighted, it's possible that scoring margin is 50 percent of the ranking process.
Basically, we went straight from margin of victory playing no part in the selection/seeding process to counting it twice, and that is producing some bizarre rankings.
Kansas has played the toughest schedule in the country without a close runner-up. The Jayhawks have taken more than their fair share of lumps on the road lately, but they have eight Quadrant 1 wins and a combined 13 Quadrant 1 and Quadrant 2 wins. In both categories, that's the most in the nation. And because of that, they are No. 1 in RPI in spite of the six losses, and they would likely be projected for a No. 1 seed under the old ranking criteria.
In the NET, however, Kansas is ranked 18th and is likely headed for a No. 3 or No. 4 seed.
The reason for that divide is margin of victory. Of Kansas' 18 wins, 11 were decided by nine points or fewer, and eight were separated by two possessions. Meanwhile, the Jayhawks had a 17-point loss at Iowa State and a 10-point loss at Texas.
They have more quality wins than anyone and have challenged themselves on a nightly basis, but the NET isn't impressed because they aren't blowing teams out.
But Kansas is going to get into the NCAA tournament regardless. It might not be fair to face a more upset-capable No. 14 seed in the first round instead of a No. 16 seed, but the Jayhawks will get their chance to play for a title.
Others might not be so lucky.
One of those teams is VCU. The Rams are No. 28 in RPI thanks to a solid showing against a nonconference strength of schedule that ranks third toughest nationally. They are 17-6 with nice wins over Texas, Temple and Hofstra and just one less-than-ideal loss to Charleston. One year ago, that type of resume would've been sitting pretty as a No. 9 or No. 10 seed.
But VCU's offensive efficiency is a disaster, ranked 223rd on KenPom. Those quality wins over Texas, Temple and Hofstra came by a combined margin of nine points, so the Rams are ranked 44th in the NET and would probably need to win the Atlantic 10 tournament to go to the Big Dance.
Similarly, Minnesota is 16-8, with seven Quadrant 1 or Quadrant 2 wins and just one "bad" loss to Boston College that bounces back and forth between Quadrant 2 and Quadrant 3. That used to be a strong tournament resume, but the Golden Gophers are ranked significantly lower in the NET (No. 58) than they are in RPI (No. 47). That's because one of the five quality wins came by a margin of more than seven points, while five of their six losses were by a double-digit margin.
And while those two teams are in danger of missing the tournament despite solid seasons, the disturbing part is the two undeserving teams that might take their place in the field.
No team has benefited more from the RPI-to-NET switch than North Carolina State.
The Wolfpack played a pathetically weak nonconference schedule, lining up eight home games against teams currently ranked outside the Top 275 on both KenPom and the NET. Combine that with a 1-6 record in Quadrant 1 games, and NC State ranks 112th in RPI.
But because the Wolfpack beat Mount St. Mary's, Maryland Eastern Shore, UNC Asheville, Maine, Saint Peter's, Western Carolina, USC Upstate and Loyola MD by a combined margin of 35.9 points per game, the NET couldn't care less about the weak schedule.
Take out those eight games, and NC State has allowed seven more points than it has scored, but it has an average scoring margin of plus-12.1 and ranks 35th in the NET because of it. That's 77 spots higher than its RPI ranking and inexplicably puts the Wolfpack in great shape for a trip to the NCAA tournament.
Nebraska is in a similar situation.
Ranked 110th in RPI with a 3-10 Big Ten record and little more than a couple of wins over bubble-y opponents, the Cornhuskers would be a complete afterthought in the bracketology conversation right now. However, four of their five best wins came by a margin of at least 15 points, and they opened the season with back-to-back wins over lowly Mississippi Valley State and Southeastern Louisiana by a combined score of 193-72.
Lo and behold, Nebraska is No. 40 in the NET and is right on the tournament bubble.
So here's the $64,000 question: Why would anyone want to put together a schedule like Kansas or VCU did when you can blow out terrible opponents for six weeks and look just as good?
That's where this move to the NET could be awful for college basketball as we know it.
Getting to the NCAA tournament is the all-important goal for coaches and athletic directors. Sure, you want to develop your players and make sure they get their degrees, but no one cares about your graduation rate if you never go dancing in March. It stands to reason that you're going to see more and more teams going the NC State route of scheduling one or two quality opponents and loading up the rest of the nonconference schedule with cream puffs.
At least when teams were "gaming" the RPI, they were filling up their schedules with borderline top-100 teams that they should be able to beat, which usually produced a lot of watchable games. They also tried to avoid Quadrant 4 games that would hurt their RPI and SOS even if they won by a million. Oklahoma's schedule for this season is a great example of the way things used to be.
But now the NCAA is encouraging teams to run up the score against terrible opponents, because if you beat Delaware State by 61 and lose to Duke by 27, good job, that's still an impressive average scoring margin.
A big reason the NCAA has been so hush hush about the NET's formula is that it is monitoring results for this season in an attempt to find the sweet spot of emphasis/weight for efficiency and scoring margin, possibly leading to changes to the formula in future seasons.
[Update: When this originally published, I poorly worded the above paragraph and made it seem like tweaks are happening throughout the season to the current formula. There are no such tweaks happening. There might be tweaks this offseason or a future offseason, as reported by CBS Sports' Matt Norlander in August, but the rating system is the same now as it was in November. I apologize for the confusion.]
By the time they "perfect" the formula and are willing to share it with the public, though, the damage may have already been done. Some coaches are already making endgame decisions based on the margin, and rest assured there are going to be a ton of feather-light nonconference schedules cobbled together for next season.
There are always going to be outliers with any ranking system, but by rating teams like NC State and Nebraska ahead of Minnesota and VCU—and by penalizing Kansas for its aggressive schedule rather than rewarding it—schools around the country are getting the message to take it easy in November and December until further notice.
We'll have to wait until Selection Sunday to see how the selection committee actually handles these extreme resumes. If they treat the NET the same way they treated the RPI and merely use it as a sorting tool before really digging into the teams and their schedules to decide who belongs, they should be able to weed out the frauds and prop up the underrated squads.
But if they still need to massage the data and seem to have just as many aberrations as before—if not more—what was the point?
Advanced stats courtesy of KenPom.com. NET rankings and Quadrant data courtesy of WarrenNolan.com. NET rankings current through start of play Feb. 10.
Kerry Miller covers men's college basketball and college football for Bleacher Report. You can follow him on Twitter, @kerrancejames.