The NET Rankings Look Like a Vast Improvement over RPI, as Evidenced by Kansas Not Being #1

Back in November, when the NCAA first released the NET rankings in its inaugural run, there were some unusual results. We wrote about it, as did plenty of others.

So in all fairness, we should write this now: the NET rankings look like a vast improvement over the RPI as we approach late January, now that more results have rolled in.

Using both the NET rankings and the RPI rankings calculated over at WarrenNolan.com here are the teams that show the biggest difference in where they are ranked, for all teams that are in the top 50 in at least one of the rankings. RPI is listed first in parentheses, followed by NET ranking for the same team.

BIGGEST DIFFERENCES, RPI versus NET in favor of RPI

Kansas (1, 19)
Oklahoma (6, 30)
Maryland (10, 20)
Nevada (11, 23)
Kansas State (13, 35)
TCU (16, 31)
Villanova (19, 28)
Florida State (22, 33)
Temple (26, 61)
VCU (30, 51)
Seton Hall (37, 56)
Alabama (39, 50)
Minnesota (40, 67)
UNCG (41, 58)
Davidson (43, 71)
Georgia State (44, 120)
New Mexico State (48, 78)
Arizona State (49, 72)

BIGGEST DIFFERENCES, RPI versus NET in favor of NET

Gonzaga (12, 3)
Kentucky (21, 8)
Virginia Tech (25, 9)
Texas Tech (28, 11)
Nebraska (47, 15)
North Carolina State (111, 27)
Wofford (45, 29)
Wisconsin (34, 16)
Indiana (51, 34)
Florida (81, 36)
San Francisco (59, 39)
Ohio State (54, 40)
Utah State (60, 41)
Murray State (87, 43)
St. Mary’s (82, 49)

Look, I am not saying that the NET ranking on Kansas is perfect either, but a big feather in its cap is that the Kansas Jayhawks are not bequeathed the #1 spot. I wrote about it four years ago when a two-loss Kansas team was ranked ahead of undefeated Duke and undefeated Virginia in the RPI in January. I wrote about it three years ago when they lost to Indiana early, but I pointed out that one could safely predict that Kansas would finish well above Indiana regardless of how the teams ended up. (Kansas could have gone 21-9 that year and still finished in the Top 10 in RPI). This story from five years ago details how Kansas gamed the RPI system in its scheduling, and they were really good at it. Kansas has been an elite program, but they’ve been just as elite at propping up their ratings by playing teams they will almost certainly beat anyway from smaller conferences but avoiding the really bad ones that sink a team’s RPI.

So far, gaming the RPI doesn’t benefit one in the NET rankings.

When you look at the teams where RPI and NET rankings differ most prominently, it’s that those handful of games against scrubs no longer kill a team’s rating. Look at NC State. They are ranked in the Top 25. But because they played two of the three worst teams in Division I (Maryland-Eastern Shore and UNC-Asheville are a combined 1-34 against other D-I teams), the RPI would be 111 right now this year. The NET seems far more practical in that regard, because it is dumb that a couple of teams you blew out in November should sink how you are viewed.

If you compare both of the ranking systems to Ken Pomeroy’s rankings, for none of the teams above is Pomeroy significantly closer to the RPI’s ranking of a team. Seventy-five percent of the teams are clearly closer in NET rating to Pomeroy’s ratings, with others being roughly in the middle (including Kansas, where Pomeroy, probably most accurately, has them at 10th, similar to where they are in the AP and Coaches Poll).

In fact, if you want to use polling as another check, it’s the RPI that is way out of whack. Gonzaga is Top 5 in the human polling and the NET. Virginia Tech and Kentucky are Top 10 teams in both the NET and polls, but outside the Top 20 in RPI. Texas Tech’s NET rating is way closer to its poll ranking than the RPI slotting. Oklahoma is incredibly 6th in the RPI (because they’ve played no teams outside the Top 200 in college basketball) but not in the Top 25 of the human polls or the NET rankings. Villanova and Nevada are the exceptions, ranked in roughly similar spots to where the RPI would be, but lower in the NET rankings.

Smaller conferences should feel good that while specific teams may go up and down, it works both way and doesn’t seem to be systematically affecting the smaller conferences. And I do also view it as a positive that Ja Morant and Murray State rate much higher in the NET than the RPI, and can at least be in the bubble conversation if the committee follows the new ratings.

I am still uneasy about the exact methodology not being transparent, and I do still have my doubts about the mentality of the selection committee and how much that will change, but the actual ratings themselves do seem like a step in the right direction.