Eli Boettger | @boettger_eli | 03/23/20
It was a relatively quiet night in the world of college basketball on Jan. 10, 2020. There were only 11 Division-I games played that evening, and the two power-conference matchups were lopsided: Iowa beat Maryland by 15 at home and Butler handled Providence on the road by 12.
Jan. 10 could wind up being an important date in uncovering the NCAA’s secretive NCAA Evaluation Tool (NET) formula, though. The NET — which is a new methodology that came into existence before the ’18-19 season — is a replacement of the outdated Ratings Percentage Index (RPI).
Just like the RPI, the NET is used as a “sorting tool” by the NCAA for its selection process. The rankings don’t suggest that every team ranked higher in NET will also be selected and seeded higher in the NCAA Tournament. It does, however, give a general assessment of team strength and is applied to the quadrant system of the same team sheets that the selection committee uses for seeding and selection.
Prior to the first-ever NET rankings release, the NCAA unveiled the infographic below, displaying the five components that make up the NET. These include the NCAA’s secretive “Team Value Index,” as well as net efficiency, winning percentage, adjusted win percentage and scoring margin. Adjusted win percentage is weighted for home and road results using the same weights as the old RPI (1.4 “wins” for a road victory, 0.6 “wins” for a home victory, etc.) and scoring margin is capped at 10 points for any double-digit margin of victory. Every overtime result is treated as either a 1-point win or 1-point loss.
This leaves us with four knowns — net efficiency, winning percentage, adjusted win percentage and scoring margin — and one unknown: Team Value Index.
Now let’s circle back to Jan. 10. NET rankings from teams ranked 11-20 are below from the morning of Jan. 10. The image is courtesy of the NCAA’s NET/RPI archives.
And here are the rankings the morning of Jan. 11:
Notice the changes. First, we have Maryland dropping from No. 12 to No. 18 after getting leveled by Iowa, which is understandable. This leads to the idle teams ranked between 13th and 18th on Jan. 10 to move up a spot in the Jan. 11 rankings.
But wait a minute: Check out Oregon and Stanford. Oregon was ranked a spot ahead of Stanford on Jan. 10 and woke up on Jan. 11 a spot behind Stanford. Interestingly enough, neither team played the night of Jan. 10.
This leads us to only one possible conclusion: One of the opponents that either Oregon or Stanford had already played had a significant result on Jan. 10. It also means that some aspect of strength of schedule or opponent winning percentage is tied into the NCAA’s secretive Team Value Index. None of the four “known” components of the NET can change without playing another game, which means that the Team Value Index must be impacted by the results of opposing teams.
Let’s take a step back for a minute and compare Oregon and Stanford at this point in the season. Obviously, these two teams must be near-identical in the NET’s measurable rating if Stanford was able to move ahead of Oregon without either team playing a game.
The chart below shows the team that had the better measure in each of the NET components as well as the stronger overall NET ranking.
This is where things get really interesting. Stanford had a higher net efficiency, winning percentage, adjusted winning percentage and adjusted scoring margin on Jan. 10 but Oregon ranked higher in NET. Clearly, the Team Value Index carries a significant weight on the overall NET rating for this to be true. In fact, if each of the four known components favor Stanford but the overall NET ranking favors Oregon, we can assume that the Team Value Index must have a weight greater than 50 percent to outweigh the remaining four components.
So how could have these teams swapped NET rankings? Let’s see if either of these teams’ previous opponents played the night of Jan. 10.
Surely enough, Butler — which defeated Stanford back in November — is the only team that either Oregon or Stanford had played that was also in action on Jan. 10. Clearly, this result was dramatic enough to impact Stanford’s NET measure to move ahead of Oregon.
Because Butler was the only team in action among these two teams’ previous opponents and neither Oregon nor Stanford played on Jan. 10, we can assume that Butler’s win over Providence is the reason why Stanford moved ahead of Oregon in the NET.
So what does this all mean? Well, it’s somewhat hard to tell at this point. What we do know, though, is that strength of opponent carries a significant weight in NET. Not only can we conclude that Team Value Index makes up at least 50 percent of the overall NET measure, but we also know that opponent strength must also make up at least 50 percent of the Team Value Index measure. This is because Stanford must have recorded a better Team Value Index rating on Jan. 11 to move ahead of Oregon, and the only result (we assume) that could have led to this change was Butler’s win over Providence. It is also plausible to believe that opponents’ opponents’ winning percentage — which was 25 percent of the RPI — could also be included in the Team Value Index.
At this point, Team Value Index seems like a variation of RPI. The NCAA shows in the above infographic that opponent, location and winner are the three factors that make up the Team Value Index, and these were the same factors that the RPI was built upon. Additionally, the old RPI was 75 percent schedule-based, and it’s clear that the Team Value Index has a heavy weight on schedule strength as well.
This study doesn’t “solve” the NET formula by any means, and it might not ever be possible to solve the formula anyways. It should, however, get us a step closer to understanding the major component that drives NET ranking. If you schedule strong teams, you will be benefitted.
Eli Boettger is a college basketball writer and founder of HeatCheckCBB.com. He has previously worked for Sporting News, DAZN and USA TODAY SMG.
Boettger’s content has been featured by Bleacher Report, NBC Sports, FiveThirtyEight, Yahoo Sports, Athletic Director University, Washington Post, Illinois Law Review and Notre Dame Law Review, among other publications. Boettger is also a current USBWA member.