Half of Basketball

I'm not a "bracketologist," which seems like both an easy and difficult thing to do. Easy because you get to make about 100 different predictions as the season goes along, and for most of those, no one cares about answers that turn out to be "wrong." Also easy because by the end of the season, all but a handful of bids have solidified into hard concrete, so there's really only 1 or 2 coinflips that have a chance of going awry. Of course, it's also difficult because, despite the Selection Committee's attempts to be as transparent as possible, the selections themselves are opaque, happening in an hours-long, closed door meeting on Selection Sunday. People do crazy things in long meetings—I once was in a meeting where someone suggested we order Chinese food. That doesn't sound crazy until you know that this was a meeting that had been going for 5 hours already, and it was also 2 A.M. Finally, I should tell you that I was the one that made the suggestion.

But the reason I'm not a bracketologist is that it's kind of like generating a "Power Rankings" list. I'm offering some ostensibly quantitative content which is actually nothing more than "that's just my opinion, man." It might have a rational basis, it might make all the sense in the world, but ultimately it doesn't help anyone understand any aspect of college basketball, and it's not very interesting other than its propensity to start threadfights on Twitter dot com.

But there are people that do practice the pseudoscience of bracketology, and those people collectively put Purdue as an 11-seed in the NCAA Tournament as of this writing. This week, the Boilers dropped a close one in Madison, and it's possible that one or more of those projections has not yet updated for that bit of news. Still, at best this is a bubble team, at worst it's on the outside looking in.

There's still time to change all of that, of course. Sweep Michigan, Indiana, Iowa, and Rutgers, and you have to like their chances (and three of those games are in West Lafayette). Still, I for one think it's a little remarkable that Purdue is even in this position. As a reminder, it wasn't just Carsen Edwards that left campus after last season. Virtually all of Purdue's outside shooting did—281 made three-pointers, which is more than any other team in the Big Ten made last year, departed. Against all odds, the Boilers' outside shooting isn't "horrible," but merely average. Still, the propensity of the players to launch 3s is well down, even accounting for the increased distance. No Big Ten team shot more 3s last year than Purdue. This year, nine do.

Purdue's defense is Big Ten average, which is a summation of the defense over the past decade or so. I think most consider Purdue to be an annually solid defensive team, but that hasn't been true in some time. Sure there, are good seasons (2017), but there are also bad ones (2014), and just weird ones (2016). Since the Baby Boilers left, it's been anyone's guess. The way in which this team defends is also constantly changing. Sometimes Purdue rebounds well, sometimes the team blocks shots, and this year the horse they ride is turnovers. If this were a much lengthier post, I'd explore my theory as to why this is—the strategy centers on the Big Guy.

And I really do think the Big Guy is at the center of Matt Painter's approach on both ends of the floor. His first Big Guy was JaJuan Johnson, who landed in late 2007. He then had one season featuring Robbie Hummel & the Pips before AJ Hammons showed up. Couple years later, Isaac Haas materializes. Then Biggie. Now, it's Trevion and Matt Haarms. And here is where we learn that 2 can sometimes be less than 1.

First, let's talk offense. Haarms is a fine player in his own right. He converts his shots, mostly (though his shooting percentage is markedly down this year). He's an adequate rebounder, and an outstanding shotblocker. But unlike the rest of the guys I listed, Haarms is not an offensive gravitational force. It is not unusual for Purdue to go a number of possessions without feeding the Dutchman. This is not the case with Trevion, who simply needs to eat. And sure, the Boilers will feed Williams on the block, who has a nice touch, but what I love about Williams' game is that he will go get the ball himself. I mean, how do you not love this?


Williams is probably open on the roll, but Reuvers does a really nice show here that puts enough doubt into Proctor that he opts to shoot the pullup instead. No matter—Williams hits Brad Davison like he's Ray Lewis and wipes him from the play. Reuvers then jumps on top of Williams, who manages to maintain his footing with a 240-pound man on his back while another Badger is hacking him. He then secures the rebound and puts it back up, with Reuvers and Wahl challenging the shot. Complain all you want about charges and and blocks and travels, but don't take this basketball away from me.

Fact is, Williams is the best offensive rebounder in the Big Ten, and one of the best in the country. His offensive rebounding has actually gotten better in conference play, so frankly, I'm inclined to get rid of that "one of" language in the previous sentence.

If Purdue didn't have Williams? Let's see, the team is 12th in conference play in 2P% (45.5). Williams leads the way at 53 percent, and Haarms and Boudreax are the only other Boilers above 42 percent in conference play. The outside shooting is average in accuracy, but below average in aggressiveness, and with the interior shooting, this nets to an eFG of 46.9 (11th). The team isn't especially great at not turning the ball over. They don't get to the foul line (13th in FTA/FGA). But Purdue is 2nd in the league in offensive rebounding percentage. And it's not hyperbole to attribute virtually all of that to Williams. In large part because he reclaims so many missed shots, he consumes over 30 percent of the Boilermakers' possessions while on the floor. This allows guys like Jahaad Proctor and Isaiah Thompson to pick their spots—those shooting percentages would be much lower if they were depended upon to take and make more shots.

Overall, the only player with a higher usage and offensive rating than Williams in Big Ten play is Luka Garza. But despite how great he's played, I think it's unlikely Williams ends up on a First Team, or even a Second. For one, he barely plays more than half of the available minutes. Unlike Caleb Swanigan, who could score from a number of spots on the floor, Williams and Haarms largely score on the block. Putting both on the floor at the same time creates too much traffic in the paint. Haarms is an exceptional shotblocker, who I'm sure Painter wants on the floor for his defense (more on that later). Of course, when that happens, it's a trade on the other end of the floor (even though Haarms is a perfectly capable offensive player in his own right—Williams is just that good).

I wanted to sum up this post with an on/off table showing the Trevion Williams Effect in conference games. And yes, we do see the effects of Williams' rebounding and 2P shooting. But there's also something curious happening...

Good Players do what now?

Purdue is actually better, in fact, much better, with Williams off the floor. The efficiency margin drops from +0.07 to -0.04 when Trevion relieves Haarms.

How can this be? Well, some of this is likely dumb luck. Neither Haarms nor Williams contribute much to the 3P game, but Williams has the misfortune of being surrounded by 29 percent 3P shooters, whereas Haarms benefits from 40 percent outside shooters. That's certainly something. But it's also impossible to miss the two-point defensive effect Matt Haarms has had. Not only do opponents shoot a paltry 41 percent on their 2s, but they attempt far more 3s—and likely not because they're open—when the Dutchman is on the floor.

On/off statistics can be tricky—but this is not a new thing. Consider last season:

Once again, a massive defensive effect that shows up in 2P defense. As detailed above, Williams is an offensive force that few teams are equipped to maintain. He put 17 on Reuvers and Company, 17 versus Watkins, 17 versus Myles Johnson, 16 on Tillman, and 36 on Jon Teske. It's hard to take away anything other than Williams is, at the very least, one of the best big men in the entire conference.

And yet, he's not the best on his own team.

And therein is Painter's problem. Haarms provides a significant defensive boost, maybe as much as +0.10 PPP on defense. But he gives back almost half of that on offense, largely because of how good Williams' offense is, and how deferential Haarms is. In the past, he's been able to put a player on the floor that was capable of being an offensive centerpiece and a defensive force. Now he has two players who are arguably individually better and offense (Williams) and defense (Haarms) than most or all of Painters' centers have been in the past. But he can only play one of them at a time, which means he's only playing one half of basketball at any given time.

Last year, Edwards and his backcourt teammates were able to cover up offensive issues with the 3P barrage, and as a result, Haarms played more than he did this season. That luxury is gone, and Painter is therefore having to make a lot of trades I'm sure he would rather not.