Q&A with Dan Hanner about Preseason Predictions

Dan Hanner released his preseason rankings from 1 (Indiana) – 345 (Grambling) yesterday. He’ll be expanding on the rankings more in College Basketball Prospectus 13 (out soon!) and ESPN The Magazine. Most of that though will probably focus on bigger schools, so I asked Dan to dive into the rankings for the conferences we cover here at Big Apple Buckets. I know lots of people want to hear about the NEC, if so, scroll a little ways down.Big Apple Buckets: Let’s start with the Ivy League. The rankings have Columbia in second and Harvard in third. (I recently wrote a piece about how people are overrating Harvard and why I think they’ll slip. Though I don’t trust the Lions and put Cornell in the No. 2 spot.) The predictions have Penn dropping to the bottom of the Ivy League. I find that difficult, but not unfathomable. I’m assuming that’s the Zack Rosen effect?

Dan Hanner: The Penn prediction almost made me fall out of my chair when I first saw it. But Penn loses its three most efficient players, Rosen, Tyler Bernardini, and Rob Belcore. I really feel sorry for Miles Cartwright now. He may never see an open shot again. And Penn isn’t the kind of program where first year players typically step in and dominate. As with most Ivy league schools, they win by developing players over time. But besides Cartwright, there just aren’t any other players you can count on for offense. Don’t be confused by Steve Rennard’s efficiency. He literally never shot last year. And while Henry Brooks and Fran Dougherty shot a lot, their efficiency was horrible. This year’s offense looks potentially historically bad for Penn.

With Harvard, this is all about Kyle Casey and Brandyn Curry withdrawing from school.

With Columbia, first they were better last year than their 4-10 Ivy league mark indicates. They lost to Penn and Harvard in OT and nearly won the other games against Penn and Harvard too. They bring back 80% of their minutes from last season. The real question is one of the forward spots. John Daniels played a lot last year and was extremely ineffective offensively. But assuming natural progressions by other players, they should get better production out of that slot. Cory Osetkowski will be able to play more minutes, and should beat Daniels 81.8 ORtg when he plays, and a healthy Skylar Scrivano might help too. If Columbia can fix that one weak link, this could be the best Columbia team in a long time.

BAB: I’m shocked to see Georgia State at two in the CAA (124th overall). Is this because Ron Hunter is considered a strong defensive coach? If that’s the case I’ll believe it, because he really turned that team around, but it seems like he lost a lot and most people expect they’ll slip. I’m also a little surprised to see Old Dominion at the top of the 2-6 group. If fact, if the 2-6 group was in reverse order I think it would match up quite nicely with how most people project the conference.

DH: The first thing I would say is that the difference between second and sixth here is miniscule. The difference between a 0.630 Pythag winning percentage and 0.555 is not much at all. Let’s assume those rankings are 100% correct, then I’m still only saying Old Dominion has a 58% chance of beating George Mason on a neutral floor. We are really splitting hairs with these teams in the middle. And the standings reflect that with all those teams at 11-7 or 10-8. Sadly this is going to be a down year for the CAA, as after Drexel, there aren’t any dominant teams in the CAA this year.

I feel comfortable with Old Dominion second. That is a large endorsement of Blaine Taylor who is clearly the best coach in the conference now that Shaka Smart is gone. ODU loses a lot again, but no one can handle it better than Taylor.

Georgia St. loses a 70% of its minutes, and so I can see why people are skeptical about this team. What the model guesses is that transfer Manny Atkins from Virginia Tech and returning team leader Devonta White give the team two offensive stars. And returning bench players like Rashaad Richardson and James Vincent should be able to succeed in expanded roles. Add in a couple of JUCO players for depth, and they aren’t going to have to rely a ton on freshmen, despite losing so much of the roster to graduation. That will help the offense and allow the team to avoid falling off a cliff. You are correct that Ron Hunter gets credit for the defense. The defense is expected to fall off from 90.6 to 95.1 because of all the new faces in the lineup, but if Hunter built a Top 20 defense last year, the model expresses some faith he can do something similar this year. (Honestly, people don’t appreciate how fantastic Georgia St.’s defense was last year. It kept them in every game. Of their eight CAA losses including the post-season tournament, seven of those were by five points or less. Six of those were by three points or less.)

George Mason doesn’t lose as much, but they lose their best player (Ryan Pearson). In this year’s version of the model, losing efficient high volume shooters (i.e. losing star players) hurts a lot because it requires other players to shoot more while getting less wide open shots. GMU also loses the very efficient Mike Morrison. So I get why people are picking them relatively high, but realistically, they have significant questions just like Georgia St. and ODU.

Delaware would be my personal pick. They got really hot at the end of last year (and the model doesn’t yet account for teams that flipped the switch late in the season). But they had enough bad losses that their stats weren’t overwhelming last season.

And Northeastern should be better, no question about it, it is just a matter of how much better.

BAB: Within the CAA, Hofstra is picked a completely respectable ninth. I was just wondering: How do the predictions see Jamal Coombs-McDaniel faring in his first season on Long Island? Where is Hofstra weak?

DH: Coombs-McDaniel is projected to be Hofstra’s best player. He was efficient at Connecticut as a sophomore and is a former Top 50 recruit. I’m predicting a 110.9 ORtg and that he will raise his shot percentage from 18% to 21%. If he doesn’t lead the team in points per game, it is only because Taran Buie decides he wants to take a bunch of dumb shots.

Hofstra’s weakness appears to be its shooting. Buie is a great scorer and going to be a key player, but he wasn’t a three point shooter at Penn St. And while Hawaii transfer Shaquille Stokes (assuming he’ll get a waiver) and Stevie Mejia are decent creators, both have struggled with shooting in their career too. All three of these guys are going to see major minutes, but their ORtgs will continue to be mediocre if they continue to be poor free throw and three point shooters.

BAB: I’m really surprised to see Siena down in eighth in the MAAC. In fact, everything made a ton of sense until I got to that spot. Why are they so low? The reliance on players that were injured / ineligible last season? The fact that they essentially played a seven-man rotation?

DH: First, I have to say, I met some of the nicest Siena fans at the Old Spice Classic four years ago. Sorry to see the program has fallen back a bit. But this prediction is pretty straightforward. They only return one efficient offensive player (O.D. Anosike) from last year. And sure, some of the players missed last season due to injury, but what can we really expect from those players this year? Rakeem Brookins had an 86.6 ORtg in his limited time two years ago, while David Martens had a 74.8 ORtg, and Trenity Burdine had an 88.2 ORtg. With stats like that when they were healthy, the upside for those players just isn’t great. There are just no guarantees in the Siena lineup, and that translates to a 6-12 team.

BAB: Alright, now let’s talk about the NEC, which I think contains some of the most interesting projections. Why is LIU Brooklyn predicted to go 12-6? I know they had a worse KenPom than Robert Morris and Wagner last season, but they return everyone but Michael Culpo and are consensus favorites to win the league title in what should be a tense battle. Here you’ve got them in a tense battle, but with St. Francis (NY) and Quinnipiac for third.

DH: You hit the nail on the head. LIU wasn’t just slightly worse than Robert Morris and Wagner last year according to the tempo free stats, they were actually quite a bit worse. And LIU was even worse than Quinnipiac.

If you compare my offensive and defensive prediction to the stats from last year, you will see that LIU is picked to improve on both ends of the court. The model certainly thinks they will be a better team this year. But the model is concerned that last year’s pace-inspired weak defense will come back to haunt the team this season.

But let me see if I can explain this a little more.

Part of the prediction for fewer wins for LIU is because other teams at the top of the NEC should be even better.

Robert Morris will clearly be better. Under Andrew Toole the offense has gone from 204th nationally to 174th to 114th, and the team brings all the key players back this year. The jump you see for them is what you expected to see for LIU and would have seen had Jim Ferry returned. Quinnipiac and Wagner’s predictions are also about what you would expect. Interestingly, the model likes Orlando Parker to have a bit of a bounce-back season and provide some much needed defensive rebounding for Wagner this season.

St. Francis (NY)’s predicted improvement is probably the one I question the most. The model likes Ben Mockford to take a step forward this year. As a former Iona recruit, his potential is still viewed as pretty strong. And the model likes Travis Nichols to step into a larger role as a three point shooter.

But the model’s result hinges on significant improvement at the point guard slot. First, Justin Newton was the part-time PG for St. Francis and his numbers were horrific. He had a 50% turnover rate and 74.3 ORtg last year. Just getting him off the court will be an improvement. Unfortunately, Brent Jones was almost equally bad and he is back. Since that was Jones’ first year season (and he made so many mistakes), he is expected to improve a lot this year. And that will matter a lot, since he was a large part of the offense.

But I will completely understand skepticism about this prediction. I’ve already written a little about how my model doesn’t give quite enough importance to PGs and it isn’t clear who will step into the part-time PG role. The JUCO addition Aleksandar Isailovic is more of a shooter, which should help the offensive efficiency, but the model basically assumes that the team’s shooting guards will split that role, and cut down on last year’s ridiculous turnovers.

The other big question in my mind is why LIU’s offensive improvement wouldn’t be higher. 107.1 to 107.2 is pretty trivial. As noted, a large part of this is the loss of Ferry. A non-trivial portion of any team’s prediction is how much player development the head coach does. The promoted assistant Jack Perri gets judged with the large number of first time head coaches across the country in the last 10 years, and their track record typically isn’t great. The model takes a more pessimistic view of his ability to bring players along than had Ferry returned.

At the player level, Jason Brickman is expected to bounce back because he was much more efficient two seasons ago. But the model fears Jamal Olaswere and C.J. Garner may have reached their potential. Olasewere and Garner made huge leaps in efficiency last year after struggling two years ago, and while both players should have similar efficiency ratings to last season, it is unlikely they will improve dramatically for a second straight year.

As I see it, the season largely comes down to how Olasewere plays. Because he shoots so often and plays such a large role in the offense, he has to improve on last year’s production again for LIU to take a substantial step forward, and the model doesn’t buy that. (Boyd is also going to be vital, but his performance has been very consistent throughout his career, so the model is more comfortable that he will continue to get better.)

BAB: Speaking of new coaches, is Wagner getting dinged for the same effect? It seems by bringing a bunch of players back – minus Tyler Murray – and adding former Michigan State recruit Dwaun Anderson they’d be able to improve quite a bit, but Robert Morris barely slips past them here.

DH: Yes, the first year coach effect is dinging Bashir Mason too. I’m not going to pretend the modeling of this is perfect, since first year coaches can be very different. But Mason’s situation is probably similar to that of Josh Pastner at Memphis. He will probably do better at recruiting initially, but it takes some time to learn how to be a winning coach.

Part of what Wagner did last year was go small, (a bit of a misnomer since the NEC isn’t the Big East) with Tyler Murray as the second tallest player on the floor. Murray was an outstanding shooter and with Murray gone, the offense is going to take a hit. But I project the team to use Orlando Parker and Naofall Folahan more which will improve the defensive rebounding and the defense overall. And that’s the projection you see, Wagner’s offense stalls out, but the defense gets better. Had Hurley returned the offensive prediction would be higher. (Also, be cautious about having too optimistic a projection for Folahan. Sure he was efficient last year, but he never shot. His passiveness means his teammates have to shoot more often and hurts the offensive balance.)

BAB: With St. Francis (NY), I actually like Mockford to take a step forward this season as well, so I guess the model and I agree there. What about the impact of Dre Calloway returning? Though, his offensive rating wasn’t great either in 2010-11.

DH: Yes Calloway is much like those three Siena players I mentioned. Here he is in year three, and we don’t really have any basis to expect a strong season. It could happen, but there is no reason to expect it.

BAB: Finally, I wonder is if LIU’s style of play – in terms of tempo and foul aversion and creation statistics have anything to do with how they beat their expectations. Messing with other teams’ rotations and things. Just something I wonder about. Might be an in-season research project with luck and seeing how it is correlated with offensive and defensive free throw rate to see if there’s any meaningful connection.

DH: Bingo: I’m projecting offensive and defensive efficiency, and assuming that predicts wins. But is that the right model for a team like LIU? I’ve long wondered whether pace shouldn’t be a larger factor here. Wisconsin typically has great Kenpom.com numbers, but they play at such a slow pace, they open themselves up to losing to inferior teams. Meanwhile LIU plays at such a fast pace, they ensure they aren’t going to lose to the teams at the bottom of the league. Also, LIU is going to take a bunch of risks if they fall behind. They are going to take chances and lose to Monmouth by 26. That could have easily been a 10 point loss if LIU had just given up and accepted it. But they took chances and gave up lay-ups to try to create chaos and get back in the game.

A big thanks to Dan for answering all these questions and make sure to check out the rankings in all the forms they’ll appear this season. It’s a bold step forward for preseason predictions.