Sweet 16 Deep Dive
And here we are, the Sweet 16. The gap between the Round of 32 and Sweet 16 always feels like an eternity to me and I can’t wait for tip-off. I watched a lot of college basketball this year…admittedly probably an unhealthy amount but it was all worth it for this. I’ve provided some of my personal thoughts from both watching these teams play this year and using KenPom’s readily available analytics.
(Like always, the code for our analysis is posted on our github giving you the ability to recreate any of the visuals)
Ken Pom Rankings
KenPom is an invaluable resource to an college basketball fan and we highly recommend his work. He provides a good entry way into more advanced analytics.
(darker = worse, lighter = better —– except for pace, darker = faster, lighter = slower)
For the first time since 2009, all 12 top three seeds advanced to the Sweet 16. The rarity of NCAA Tournament “chalk” results also leaves 14 of the 16 four seeds or lower alive, tying a 2009 record. The results play well to the committee’s intended matchups for the second weekend of the tournament, matchups that present great intrigue. However, in a tournament that generally puts together teams with vastly different styles of play, this Round of 16 does almost the opposite.
Consider this: Two matchups—(2) Kentucky v. (3) Houston and (1) Virginia v. (12) Oregon—involve teams within a quarter of expected possessions per 40 minutes, with all four teams in the bottom third slowest teams in the nation (first to 50 points wins?). More staggeringly, all remaining teams rank in the top 50 in defensive efficiency (out of the 353 teams in Division 1) and all matchups feature teams within 40 slots from each other.
Digging even deeper, six of the eight games—all but (2) Michigan State v. (3) LSU and (1) North Carolina v. (5) Auburn—showcase defensive efficiencies within thirteen slots of each other, and five present a matchup where both teams rank in the Top 20, including the two highest rated defensive teams in the nation—(2) Michigan v. (3) Texas Tech—in a single matchup.
While most strong defensive efforts correlate directly with slower tempo (nine of the sixteen teams have a tempo in the slowest half of the country—eleven if you include two teams just above half), four teams are included in the 70 fastest teams in the nation (North Carolina, Duke, LSU, Gonzaga), with North Carolina ranked sixth.
Half of the matchups involve a difference in speed of about 100 spots, which is a handful of possessions per game, but we have only one Top 100 team going against a team ranked 300 or lower (Duke against Virginia Tech). In fact, even with the larger differentials, three of the four matchups pit teams together that are in the same speed “bracket”, or quadrant of teams within an acceptable margin from the mean. Simply put, almost every team playing each other in the Sweet 16 plays at a comparable speed and all with a high level defense.
Net Ratings
Net Rating is a metric encapsulating a team’s offensive and defensive ratings. When looking at the distribution curves for each team, a wider curve, like Oregon’s curve above, means they have much higher variation from game to game. They could play amazingly one night and terribly the next. A tighter curve, like Houston’s, means they are more consistent.
Offense
As expected, the sweet 16 teams are loaded with highly efficient offenses and the matchups are even more incredibly balanced, with only Virginia v. Oregon presenting a significant difference in offensive production. Six out of the remaining seven matchups are within ten positions on the offensive efficiency rankings. It’s noteworthy that several matchups employ similar offensive schemes or personnel, none more eye-opening than Gonzaga v. Florida State, which we’ll discuss in a minute.
So we know how the field shakes up, but what about each individual game? Let’s take a look at each matchup complete with A.I. Sports’ official model win percentages. As far as the score predictions go, let’s just say they are half computer, half man and we do not recommend betting based off of the scores (especially the scores that don’t match our official models).
Most Intriguing Games of the Sweet 16
Michigan – Texas Tech
To me (and the model), this is by far the most intriguing game of the Sweet 16. The two best defensive teams in the nation. Rising NBA talent Jarrett Culver vs. Michigan’s Charles Matthews. Texas Tech’s ferocity on offensive boards. Michigan’s remarkably steady defensive setup. The 5+ blocks a game for Tech. Michigan’s composure during a grueling schedule and 2018-Los Angeles-esque “home team advantage.”
On the surface it looks like this would come down to who has the better offense, and on paper that is Michigan, but Tech has done an excellent job limiting 3 point shots, something Michigan has rode the last couple years deep into the tournament. There are concerns that if Culver and Davide Moretti are unable to score for Texas Tech, the offense won’t be there, but with one of the best offensive rebounding schemes in America, the Red Raiders have the ability to create shots out of nothing.
If Matt Mooney can get as hot as Jordan Poole was last year for the Wolverines, the defensive strength down low for Texas Tech may be too tough for Michigan to handle. Because both teams excel defensively, I’m extremely interested to see how hard both teams have to work to score. The scoring certainly comes more naturally for Michigan, but they’ve been known to go silent for games at a time.
I can’t wait for this matchup. The tournament has a funny way of finding new stars and by tournament’s end, I think Culver plays his way into a top 10 pick in the NBA draft. Neither team eclipses 65.
Model: Michigan 73%
Monte Carlo: Michigan 72%
Tennessee – Purdue
Tennessee’s strength is their deep guard rotation. Purdue’s strength is a star guard. Can one player make up for depth issues?” Tennessee’s offense runs through the scintillating play of Junior Guard Jordan Bone. His ability to distribute without turning the ball over has him at a school record 2.72 assist-to-turnover ratio, leading to a Top Ten shooting efficiency for the Vols. Tennessee also likes to run when they can, leading to a Top Ten scoring output of over 80 points a game.
But don’t get it twisted, Tennessee is an excellent half-court offense and can pick and choose when to attack from their sets. With a number of options during each possession including Senior Admiral Schofield, Redshirt Junior Lamonte Turner, and Tennessee’s best player, Forward Grant Williams, Purdue will have to work extremely hard each possession to keep Tennessee off-balance. However, 7’3” Matt Haarms may actually force Tennessee to go away from Williams, and force the guards to make shots, something Purdue has been excellent at denying over the last month (Purdue held teams in the mid-30s in field goal percentage and high 20s in three-point percentage during that span). But Haarms has only averaged about 20 minutes a game and Purdue isn’t very big otherwise, so look for Tennessee to play Williams big minutes to take advantage of when he is out.
While Tennessee may have a counter for Purdue’s Haarms, they may not have any way of stopping the hottest player in the country, 6’1” Junior Guard Carsen Edwards. Edwards dropped 42 points against defending champions Villanova, the most points in an NCAA Tournament game since 43 by Gerry McNamara. Tennessee is one of the nation’s oldest and most experienced teams, while Purdue lost four players from last year’s Sweet 16 team. This game is all about individual matchups, and Purdue will have to do their best to play Haarms’ minutes intelligently, while leaning on the scoring of Edwards and 6’6” Senior Guard Ryan Cline. These are two of the smartest teams in the nation and it should be extremely close.
Model: Tennessee 53%
Monte Carlo: Tennessee 54%
Mirror Matchups:
Virginia – Oregon
Defense and 3 point shooting. Defense and 3 point shooting. The recipe is the same for both teams. Stifling defense, active hands, outside shooting, and deliberate (read: prodding) offensive execution. Although Oregon is the only double-digit seed remaining, they are a Virginia Lite. After losing freshman Phenom Bol Bol for the season, Oregon had to reestablish themselves as a perimeter oriented team with guard heavy lineups and outside shooting. While their inside scoring is greatly impacted without Bol, they have still shown a penchant for blocked shots and can run in transition when needed. Virginia, however, is an extremely tough matchup for Oregon. While Oregon thrives on ball pressure and turnovers, Virginia just flat doesn’t turn it over. And while Oregon tries to three-point shoot you out of the gym, you guessed it, Virginia just flat doesn’t let you. Virginia is extremely comfortable playing a game in the 50’s and with four capable guards, a Top Ten NBA prospect in DeAndre Hunter, and a newly rejuvenated inside force in Mamadi Diakite, they finally have the offensive firepower to pull away. The matchup to watch is Virginia point guard Ty Jerome (rapidly ascending NBA draft boards) and Oregon point guard Payton Pritchard. This may be the best point guard matchup in the entire tournament. Popcorn ready.
Model: Virginia 65%
Monte Carlo: Virginia 66%
Gonzaga – Florida State
The best offensive team in the nation has revenge on its mind. After being a tournament favorite in 2017-18, Gonzaga ran into its nightmare in the Sweet 16: Florida State. Gonzaga’s offensive prowess lies on the broad shoulders of NBA Draft fast-riser Rui Hachimura, Center Killian Tillie, and Forward Brandon Clarke, a front line that can rarely be matched across the country. But Florida State may be the only team who can match up size-wise with Gonzaga. With their own mountains in 7’4” Christ Koumadje, 6’10” ACC 6th Man of the Year Mfiondu Kabengele, and 6’8” Phil Cofer—it must be noted that Cofer is hobbled by an ankle injury and is mourning the passing of his father late last week—FSU presents the exact kind of matchup problem Gonzaga was hoping to avoid.
Throughout the season, Gonzaga has been the epitome of efficiency, topping the nation in field goal percentage and assist-to-turnover ratio. Yet their offensive prowess sometimes overshadowed an incredibly stout defense, a defense that boasts a Top Ten rating in field goal percentage defense and blocks per game. Florida State is coming into this matchup hot and confident, winning 16 of their last 18 games, with their only two losses coming against Duke and UNC. Florida State’s game-plan is simple: shoot threes, get to the paint. Although they can play a bit out of control at times, their goal is to use their paint prowess to squeeze in the defense and dish it out to their long and athletic guards, which include Terrance Mann and MJ Walker. The guards have the option to shoot the three or drive to the newly vacated space at the foul line.
To me, this matchup comes down to Terrance Mann vs. Gonzaga’s Zach Norvell Jr., two players who can get hot at any time. I have trouble ignoring the revenge factor in this one, amplifying the eternal chip the WCC’s Gonzaga has on their small conference shoulders. And to be fair, Gonzaga’s most recent loss to St. Mary’s was against a team that plays nothing like Florida State. I can see this game going down to the wire, but the first ten minutes will be key. The Noles jumped out to a big lead last year early but if model-favorite Gonzaga can play their game and limit turnovers, they should be able to pull this one out.
Model: Gonzaga 91%
Monte Carlo: Gonzaga 81%
The Rematch
Duke – Virginia Tech
Duke wants to run. VT wants to crawl. In a rematch of a 77-72 VT win at Cassell Coliseum in late February, a place where Duke hasn’t won since 2015, most will caution the VT hype with one simple word: ZION. Zion missed the ACC clash after the infamous shoe incident. VT took advantage. While only scoring 26 times in the paint (to Duke’s 24), Tech got to the line 29 times to Duke’s 19. After the game, Coach K said, “[Duke] didn’t lose tonight. Virginia Tech won. There’s a big difference.” So with Zion back, Duke should breeze, right?
Maybe, but the numbers caution you against going all in on Duke. Justin Robinson, VT’s senior point guard was one of the most efficient players in the nation before going out for two months prior to returning for the first round matchup against Saint Louis. With Robinson, Virginia Tech was a top five team in three-point percentage nationally; without him, they struggled mightily, including going 3-28 in a February loss to Virginia. Duke, on the other hand, has struggled for most of the year from three.
Virginia Tech loves to bait teams into bad shots with their number of defensive looks, including an extremely athletic matchup zone led by NBA prospect Nickeil Alexander-Walker. But Duke is Duke and Duke has four of the best players in the nation, who, when the game-plan doesn’t work, can iso you with ease. If RJ Barrett plays the way he did against Virginia Tech in February and Zion is even just a shell of himself, Duke should be able to navigate Tech to a win. But this will not be easy, and the familiarity in this matchup will keep Duke fans sweating until the end.
Model: Duke 93%
Monte Carlo: Duke 84%
Who is Luckier:
Kentucky – Houston
This matchup is by far the hardest to read. Will Kentucky stud PJ Washington play? Has Houston tested themselves enough this year to be able to go toe-to-toe with Kentucky (their 42nd ranked strength of schedule has been ballooned a bit by wins over Oregon, LSU, Utah State, St. Louis, splitting season series with UCF, and winning 2/3 from Cincinnati). According to the Pomeroy Index, these are two of the luckiest teams* remaining in the tournament, pointing to their wins being a bit inflated compared to their efficiency rankings.
However, these two teams are still extremely good. They are active defensively (they both average almost five blocks and over six steals a game), they are both very deliberate offensively, and they can both score. While their tempos are in the bottom slowest in the country, Houston likes to run to set up offense, while Kentucky has the skill, specifically with Point Guard Ashton Hagans, to force turnovers and run.
Expect Hagans to spend some time checking Houston Guard Corey Davis Jr. who hit seven threes against Georgia State and finished with 21 points against Ohio State. And look for Houston to play a deeper bench than Kentucky, who, without PJ Washington, have given big minutes to Forward Reid Travis, and the Guard trio of Tyler Herro, Ashton Hagans, and Keldon Johnson.
Athletically, both teams are very similar, and it will come down to offensive execution and three point shooting, something Kentucky has been hit or miss on this year. Kentucky is 13-1 all-time against Houston with their only loss coming in 2017. With the game being played in Kansas City, expect both teams to have large fan bases. I can see this game going either way, but I have a hunch that PJ Washington plays for Kentucky, and he will be the difference in a tight matchup.
Model: Kentucky 55%
Monte Carlo: Kentucky 54%
The Numbers Say… Not Close:
North Carolina – Auburn
North Carolina does what Auburn does. Unfortunately, North Carolina does it a bit better and more frequently than Auburn. Carolina will be running from the tip, something that the athletic Auburn should be able to handle, but their half-court defense, which surprisingly has been the source of the majority of their turnovers caused—they do an excellent job of cutting passing lanes with their length on the wings—should be neutralized by how quickly Carolina gets into their offense.
UNC’s athleticism should also give Auburn some problems in the half court, but one potential area of exploitation for Auburn is Carolina’s mediocre three-point defense, (ranking in the lower half of the ACC during the year) as Auburn has made at least 12 threes a game during their most recent winning streak. Auburn has seen some of UNC’s athleticism this year (most notably from LSU, Kentucky, and Duke), but the consistency of defensive effort and positioning hasn’t always been there from Auburn and after watching Kansas’ implosion, Carolina will be ready to neutralize the chaos. Plus, tired legs against UNC (Auburn has played seven times in two weeks) is never a recipe for success.
Model: North Carolina 63%
Monte Carlo: North Carolina 61%
Michigan State – LSU
Simply put, Michigan State is a better team. However, if there is one area LSU can salivate over, it is Sparty’s penchant for turnovers. They committed 22 turnovers (leading to 26 points) against Minnesota in the Round of 32. That simply cannot be overcome as they get deeper into the tournament. Although Michigan State has had problems with turnovers all season, they have been fairly consistently good in that department over the last month, so there is reason to be optimistic that it won’t continue, especially considering how well they dominated Minnesota.
Many would argue that Michigan State has the best player in the matchup in six-foot Junior Guard Cassius Winston, but I’d argue he hasn’t seen a player as dynamic as LSU’s Sophomore, 1000-point scorer Tremont Waters. Waters has an uncanny ability to get to the rim and his leadership and poise saved LSU in close games against Yale and Maryland.
LSU loves to clog the lane and block shots, and fast riser Naz Reid and Kavell Bigby-Williams make it very difficult to score inside. But both players tend to overcommit to block shots, which leaves back-cuts open for good finishers. LSU’s poor three-point shooting means they can go scoreless for minutes at a time, and with a defense as strong as Michigan State’s they should be able to take advantage during those periods.
If Nick Ward is back to full strength, MSU should be able to rely on the quartet of Xavier Tillman, Kenny Goins, Aaron Henry, and Ward to counteract LSU’s bigs. LSU needs to play with controlled pace and their biggest chance will rely on Michigan State turning it over. The differential is just too big, however, and I believe Michigan State won’t struggle too much with the athletic ability of LSU. Close early, pull away late.
Model: Michigan State 74%
Monte Carlo: Michigan State 75%
Elite 8 Predictions
Duke – Michigan State
Gonzaga – Michigan
Virginia – Tennessee
North Carolina – Kentucky
*A quick disclaimer on KenPom’s “luck” index. Teams with 1 or 2 very good to elite free throw shooters tend to be considered lucky, particuarly solid but not great teams. That is partially because those teams fare well in close games. However, anyone who watches enough basketball knows that more often then not, when you hit your free throws, you win the game. There are certainly some valid aspects to the computation but we would be remiss not to mention its shortcomings.
