I’ll believe the NCAA ditching RPI is a meaningful change when I see it

Since the universe hates me, two rather significant items of basketball news dropped yesterday while I was in court. (Luckily, my co-bloggers picked up on both.) First, Vanderbilt dropped its non-conference schedule for the 2018-19 season. And second, the NCAA announced that it was ditching the RPI in favor of the new NET rating.

I’ll have more thoughts on the former closer to the season, but my initial thought is that the concerns about overscheduling from last year are mostly gone. With nine games at Memorial Gym and basically all of those being against beatable teams, but with enough tough road games to impress the Selection Committee, this is fine. As for the latter...

I’ll defer to literally a rocket scientist on the math behind the new formula. The details on the formula are a little sketchy, but the new formula will apparently take margin of victory into account, albeit with a 10-point cap. That’s dangerous, as a 10-point cap basically neuters any useful margin of victory-based system while also potentially putting an inordinate amount of stock on whether you make your free throws while the other team is trying to extend the game in the final minute. (On the other hand, if this stops coaches from fouling while down four with 20 seconds left, I am here for this.)

And as many issues as the RPI had, one of the nice things about it was that it was really damn transparent. Everybody knew what went into the RPI, to the point that multiple outlets could reproduce the RPI accurately. On the other hand, nobody knows what exactly goes into KenPom because he uses a proprietary formula, and the NCAA hasn’t completely divulged the formula, and probably won’t.

The real issue that people had with the RPI, though, was that it was so transparent that coaches had figured out ways to manipulate the RPI through, basically, scheduling smart. For a potential NCAA Tournament team, playing a bunch of teams ranked around 150 instead of around 300 in your nonconference schedule made a significant difference in your RPI and strength of schedule numbers. There’s a good reason for that: the #300 team is not beating the #50 team, no matter what, but the #150 team might beat you if you’re having a bad night. But racking up a bunch of wins over mediocre teams rather than terrible teams was, apparently, a negative.

With all that said, I’m not convinced that this change will make much of a difference.

The dirty secret about the RPI is that the NCAA didn’t really appear to rely on the RPI all that much. Oh, sure, when a team’s RPI was really bad, they weren’t getting selected for the NCAA Tournament, and a top-25 RPI team getting snubbed was virtually unheard of. On the flipside, it’s only been three years since a 27-6 team with a top-30 RPI (but a KenPom rating of 68) was left out of the tournament, and it’s only been five months since the NCAA gave an 8-seed to the #69 team in the RPI and stuck the #21 team in the RPI in the play-in game (and left out the #34 team in the RPI — haha, screw you, Middle.)

There are lots of other factors the Selection Committee took into account, like records vs. top 25, top 50, and strength of schedule, but over the years the reality has more and more looked like the Selection Committee selecting and seeding the field however they feel like and then using the RPI after the fact to perform the necessary mental gymnastics to explain its decisions. Earlier this year, neither RPI nor KenPom would have explained Middle being out of the tournament while Alabama was a 9-seed, for instance.

If the NCAA had simply selected the 36 highest-rated teams in the RPI that didn’t earn automatic bids, and were now selecting the 36 highest-rated teams in NET that don’t earn automatic bids, this would be a meaningful change. As it stands, though, the only thing that will change is what rating system the NCAA uses to explain away its more controversial selections.