clock menu more-arrow no yes

Filed under:

Rutgers Men’s Basketball vs. The Computers

New, comments

After three incredible performances in a row, Rutgers men’s hoops finds itself somewhere in the NCAA bubble conversation. Here we talk about the computer rankings and how to make sense of it all.

NCAA Basketball: Rutgers at Wisconsin Mary Langenfeld-USA TODAY Sports

After a dominating and incredible 73-65 road win at Wisconsin, Rutgers men’s basketball now stands at 15-9, 9-5 in Big Ten conference play. What a remarkable turnaround from the Scarlet Knights!

In my last KenPom update, I wrote in part:

Realistically, unless something weird happens like winning the Big Ten Tournament (or winning several games as a moderate-to-heavy underdog in February), the ceiling for this team is probably now a home game in the first round of the NIT. (Which, by the way, none of us should scoff at. I’d be there screaming my lungs out. Historically, this alone is a well above average season for Rutgers hoops.) That said, because of how this schedule is backloaded, we won’t know until end of February whether Rutgers is one of those strange teams that plays to the level of their competition. With six consecutive ranked opponents after Northwestern, we’ll know after a few more games, so stay tuned.

Well, something weird (in a good way) certainly happened. The big debate around these parts is Rutgers vs. the computer rankings, and as OTB’s resident stats nerd, I thought I might weigh in with a bit of context. Here’s where Rutgers currently stands after yesterday’s win in Madison:

  • KenPom: 84 to 78 (was 104 prior to the Michigan State win)
  • T-Ranketology: 0.8% chance of a bid to 1.3% chance
  • NET: 94 to 81 (was 104 prior to the Michigan State win)

Hmm. None of these numbers really inspire confidence, especially when you consider how much it seems the Selection Committee thinks they’re important. (Though it’s also not entirely clear how the Committee weights the eye test in contrast to the metrics, because by the eye test and the Quad records, Rutgers looks way better — Rutgers already has five Quad 1 victories with roughly five more Quad 1 opportunities left on the schedule.)

Overall, they are indicative of the larger problem, which is that the computer models have no real way of handling a team like Rutgers, which had a terrible November but is having a transcendent February so far.

If you were to only look at the above numbers, Rutgers is still a team on the bad side of the NCAA bubble right now, even after three impressive wins. How impressive? Rutgers’ last three wins had the following pre-game expectancies (per KenPom): 31%, 38%, 23%. On pure chance alone, a team is expected to win those three games as an underdog only 2.7% of the time. So something’s going on with Rutgers basketball that’s really positive and seems to be enduring.

We’ll break down the narrative next, but before we do that, let’s get New Wave in here…

Breaking Down the Narrative

Rutgers fans know the story by now. There have been three really bad losses in total. You can sort of forgive the bad losses at DePaul (Ranked #102 in KenPom, though technically a Quad 2 loss), and at UMass (Ranked #181 in KenPom). If that were all, I’d be writing a different article, but Rutgers also lost at home to Lafayette (#322 in KenPom), a team with seven D-I wins (8-15 overall record).

In short, November was a terrible, no-good, very bad month for Rutgers basketball, but how do you juxtapose this with what might be the most impressive February in the making? This is what the computer models are struggling with. They’re designed to handle a certain constrained set of priors, and Rutgers this year is throwing the damn system for a loop.

In situations like this, I look to comps, but there really aren’t any for this Rutgers team in the brief NET era (it took over for RPI for the 2018-19 season as the primary quantitative evaluation tool for the Selection Committee). The closest two comparisons probably are:

  • 2019-20 Kentucky, which was ranked #2 in KenPom when it lost at home to #294 Evansville. (The difference in KenPom ratings for this home loss was 290, compared to a difference of 243 for the then-rankings of Rutgers and Lafayette when they squared off earlier this season.) That loss got lots of news at the time, but it dropped Kentucky all the way to… #5 in the KenPom rankings. Kentucky ended up not a top-five team that season, finishing the year at 29 (one spot lower than Rutgers) in the KenPom rankings. Had there been a 2019-20 NCAA Tournament, you could safely assume both Kentucky and RU would have been in that ~9 seed range. Of course, it helped that Kentucky avoided any other really poor losses that season, whereas Rutgers this season has two on top of the Lafayette loss.
  • A second comp, which Rutgers fans might chuckle at, would be 2018-19 Seton Hall. SHU finished that season 20-13, but with three Quad 3 home losses (Xavier, DePaul, Saint Louis) and having had two separate 3+ game conference losing streaks. Nonetheless, in a relatively weak year for the Big East (SHU was only one of three bids that season for the conference), they made it into the tournament as a 10 seed. But in this instance, none of SHU’s home losses that season were to a Lafayette-caliber team (then again, they only had five Quad 1 wins all season, Rutgers this season already has six). This is kind of a comp, but again, not perfect.

So where does this leave us?

Given there’s never been a team like Rutgers in the history of the NET ranking, it’s not surprising the metrics don’t do Rutgers justice. Right now, the whole situation leaves us in a place where Rutgers is on the bubble list, but not yet on the good side of it. What needs to happen for Rutgers to make it to the NCAA? Here’s my take.

Rutgers could go ahead and leave no doubt by winning their next two games, vs. Illinois (35% win expectancy) and against Purdue at Mackey (12%). If Rutgers did this, they’d have five consecutive ranked Quad 1 victories, eight Quad 1 wins with 3-4 more opportunities remaining, and they’d be 11-5 in Big Ten conference play and probably nationally ranked in the AP Top 25. Again, if this happened, their KenPom rankings would improve to the point where they’d be toss-ups at least in every remaining game on their schedule.

But the above scenario probably won’t happen. More realistically, Rutgers just needs to win four more games between now and Selection Sunday to make them an at-large bid. To be clear, this could be 3 home wins + 1 conference tournament win. It could be 3 home wins, 1 away win, and a close loss in the first conference tournament game. It could probably even be 2 home wins + 2 conference tournament wins. But, to me, four is the magic number.

(I know some other people think three is the magic number, but I do feel like those people are ignoring the qualitative piece of the committee’s selection process. Some of the teams above Rutgers on the bubble right now include North Carolina, Florida, Michigan, and Syracuse. If you’re the committee and you have no attachment whatsoever to Rutgers, and you had to flip a weighted coin, are you going to pick Rutgers or a blue blood? I’m not trying to be a doomer, and I usually hate the “RU Screw” narrative in general, but here I think it makes at least some sense. Rutgers probably needs to be qualitatively better than the above mentioned four teams, plus maybe a few others, in order to squeak into the NCAAs.)

The look ahead

The schedule does not get easier for Rutgers anytime soon. Here is the remaining schedule for RU, this time in order to their likelihood to win the game:

  • March 6, vs. Penn St. (65% win expectancy)
  • February 26, vs. Wisconsin (49% win expectancy)
  • February 16, vs. Illinois (35% win expectancy)
  • March 2, at Indiana (29% win expectancy)
  • February 23, at Michigan (27% win expectancy)
  • February 20, at Purdue (12% win expectancy)

KenPom predicts Rutgers will win two of the above six games, finishing the regular season at 17-13, 11-9 in conference play. But, and I can’t believe I am saying this given who I am as a person, I don’t believe this projection because it’s based on an outdated expectation of how Rutgers basketball is playing. You have to somehow remove the November data to get a sense of what this team is all about, and though I can’t do that with my KenPom account, I’m pretty confident if this were possible, you’d find 3+ wins out of the next six games. Even if KenPom’s projection holds true, Rutgers should finish the regular season with 7+ Quad 1 wins, which despite what I wrote earlier might be enough to tilt the committee in their favor. Opportunities clearly exist for this Rutgers team.

I was pessimistic about this team’s chances in my last update, but I’ve completely changed my mind. Rutgers is trending upward in a way I’ve never really seen any team trend upward before, and all that’s left to do is hope the upward trend continues for a few more weeks.