UVA Basketball Preview: Can History Help Predict How Good the Hoos Will Be In 2019?

By: Christian Rogers

Last year at this time, Wahoo basketball fans were nervously excited for the season, desperately hoping that this would finally be the year. The team was coming off the most devastating loss in program (and possibly NCAA) history, and many feared that another core group of studs would leave school empty-handed. This year, it’s a little different. In April, the Hoos became the first first-time champion since Florida in 2006, and they did so in a historic manner. Everyone knows how close the final three games of the tournament were, but just how crazy was the run to the championship? Per Kenpom’s minimum win probability data (i.e., the lowest chance of winning the team had at any point in the game) the Hoos had the most improbable tourney ending of the past ten years in terms of how the games played out:

Screen Shot 2019-11-06 at 9.15.58 AM.png
As the graph shows, the Wahoos were the only champion with a minimum win probability under 15% in each of their final three games. In 2014, UConn was the only other team with two such games. Some may view this as demonstrating that the Hoos were lucky to win the tournament. While there’s no denying that good fortune played a role, it’s also fair to conclude that Tony Bennett absolutely deserved to win a national championship, given the incredible program that he has built in Charlottesville. Bennett took over a program in shambles in 2009 and has turned it into a powerhouse. Just how good have the Hoos been during his tenure? Here are the team’s adjusted offensive and defensive Kenpom ranks in the Bennett era:

Screen Shot 2019-11-06 at 9.17.01 AM.png

While the offense has been up and down, the Wahoo defense has ranked top 10 every year since 2014, ranking outside of the top 5 just once during this time frame. Unsurprisingly, the Hoos are the only program in the country with this distinction. It’s probably a given that the team will once again rank as one of the best defensive teams in the country – at this point, it’s become as surprising as Alabama being good at football. 

DeAndre Hunter’s entrance into the NBA was expected after this season, but the additional departures of Kyle Guy and Ty Jerome mean that Coach Bennett will likely have to rely on new faces more than he is used to. The NBA trio, as well as graduated Jack Salt and other departures like Marco Anthony and Francesco Badocchi, combined for roughly 60% of UVA’s minutes last year, scored just under 70% of the team’s points and made over 75% of the team’s three-point shots. This amount of roster turnover, while novel to the Wahoos, is nothing out of the ordinary for reigning national champions. In fact, for the past 17 champions, Virginia’s returning minutes and points percentages are right in the middle of the pack:

Screen Shot 2019-11-06 at 9.18.15 AMCan the performance of these past champions based on their returning minutes and points tell us anything about how the Wahoos will perform in the upcoming season? To determine the quality of a team, I used Kenpom’s adjusted efficiency margin, which is the number of points by which a team would be expected to outscore an average Division I team over 100 possessions (full explanation here). This metric is an easy way to compare team quality from year-to-year and uses a very simple unit (points per 100 possessions). The variable that I looked to predict was the increase or decrease in the team’s adjusted efficiency margin the year after winning the national championship. Aside from the returning minutes and returning points of a team, another thing considered was whether recruiting had any impact on a team’s performance as a defending champion. Virginia is no Duke or Kentucky, teams that routinely bring in multiple 5-star recruits (in fact, Coach Bennett has never had a one-and-done recruit). To find recruiting data, I used 247’s national team rankings, which stretch back to 2003. 

Due to a combination of talent leaving and natural regression to the mean, every defending champion beside one got worse the following season. The exception was 2006 Florida, which improved by 2.53 points per 100 possessions and won the championship again. The biggest drop was 2012 Kentucky, which got 19.34 points per 100 possessions worse, after dominating the country with Anthony Davis and company the year before. Here are scatterplots for all three variables versus change in efficiency margin:

Screen Shot 2019-11-06 at 9.19.03 AM.png
These plots demonstrate a clear positive relationship between efficiency margin change and returning minutes and returning points. This makes sense intuitively, as we expect returning players from the defending champions to be good, and for a team with more of these returning players to be better than one with less. In fact, the correlation between efficiency margin change and returning minutes percentage is 0.83, and for points, it is 0.85. Both of these numbers indicate a strong relationship between the variables. In this case, the anecdotal evidence is to support the importance of continuity is also strong. Take 2006-2007 Florida, the only repeat champion this century. They featured future All-NBA players Al Horford and Joakim Noah and brought back an astounding 90.3% of its minutes and 93.2% of its points. The only other champions to bring back more than 70% of both minutes and points were 2013 Louisville and 2016 Villanova, who were second and third best in terms of efficiency margin change.

On the flip side, no significant relationship appears to exist between efficiency margin change and recruiting rank. The correlation between these variables is a measly 0.33. This does not seem to make as much sense, as we’d expect a defending champion bringing in top-notch recruits to maintain a certain level of play. However, looking at some of the anecdotal evidence, you can start to see why the relationship is very weak. For example, Kentucky’s 2011-2012 team won 38 games and featured the #1 and #2  NBA draft picks (Anthony Davis and Michael Kidd-Gilchrist). The following year, Coach Cal reloaded with the number 1 recruiting class in the country, led by Nerlens Noel, the #1 overall recruit. Yet, the team won just 21 games and didn’t even make the tournament. Noel did only play 24 games that season due to injury, and the team went 17-7 with him on the court. However, that included 8 wins over non-power five competition, meaning they were 9-7 against better competition, hardly the mark of a powerhouse. Another example is Duke’s 2014-2015 team, which won 35 games, and was led by future first-round picks Jahlil Okafor, Justice Winslow, and Tyus Jones. That offseason, Coach K brought in the number 2 recruiting class in the country, headlined by future second overall pick Brandon Ingram. We’d expect this team to still be really good (and they were picked #5 in the preseason poll), but they won 25 games and only got a 4 seed in the tournament, the worst seed that coach K had received in the tourney since 2007. 

Why didn’t these two teams (and others) fare better with such highly regarded recruits? The main reason is that it’s rare that even top-notch recruits become dominant college players right away. Any team that just won a national championship is going to lose excellent players, and replacing them is not as easy as it may seem, even for a blue blood like Duke or Kentucky. Consider this plot:

Screen Shot 2019-11-06 at 9.20.03 AM

This graph demonstrates the number of top ten and top 50 recruits who finished in the top 100 in Box Plus-Minus in their freshman season. Box Plus-Minus is a simple metric that uses box-score statistics to estimate how many points a player provided his team above an average player per 100 possessions. Therefore, the top 100 in this stat is a basic proxy for the top 100 most valuable players in the country in a given year. As this graph shows, a very small percentage of the top 50 recruits are high impact players right away. Top ten recruits fare better, but even for them, the chance is a little worse than a coin flip. Sometimes, a team is fortunate and gets multiple high-impact recruits (take Kentucky in 2012 with Davis and Kidd-Gilchrist). However, the odds of this happening two years in a row are slim; just because one group of recruits succeeded with a coach doesn’t guarantee success for any future classes. Since the success of recruits is fairly random and extremely hard to predict, recruiting does not offer much insight into how a champion will fare in the following season.

I also repeated the analysis with teams that were great in general (which I will call “top teams”), because the tournament can be very fluky and the best team doesn’t always win it all. I defined this as teams greater than 30 points per 100 possessions in adjusted efficiency margin, as this level is usually only breached 1-2 times per season. From 2004-2018, there were 27 such teams. The results were very similar to the analysis conducted for national champions. Returning minutes had a correlation of 0.84, while points’ correlation was 0.83, indicating that the relationship holds even when extended outside of national champions. During this time frame, only 3 “top teams” got better the following year: 2008 North Carolina, 2014 Arizona, and 2015 Villanova (UNC and Villanova went on to win national championships the following year). All three returned greater than 55% of their minutes and points the following season. To sum it up, the success of great teams and national champions the following year is heavily dependent on how much continuity they have from that season. Bringing in highly-rated recruits doesn’t appear to determine much, based on how unlikely it is that even the best recruits in the country have the same impact as great players that left school or graduated.

What does all of this mean for the Hoos in the 2019-2020 season? The team will bring back 40.1% of its minutes and 32.4% of its points, numbers that are in the middle of the pack for past national champions. Therefore, we should expect to Hoos to be worse in 2019-20, but not to the point where we should worry about making the NCAA tournament. We can get more specific, however. I created linear regression models for each of the four variables (returning minutes and returning points for champions and “top teams”). Returning minutes and returning points could not be included in the same model, as these two variables are highly correlated with each other. 

With these models, I calculated 95% confidence intervals for Virginia’s change in efficiency margin for the upcoming season. These ranges will give us a good idea of where the Hoos should end up, based on historical results of national champions and other top teams. The average range of the four confidence intervals was a decrease in efficiency margin between 11.96 and 8.46 points per 100 possessions, with a midpoint of a 10.2 point drop. Last season, the Hoos had an adjusted efficiency margin of 34.22 points, meaning the confidence intervals would place them between 22.26 and 25.76 points better than an average team per 100 possessions, which is still extremely good. The low end of 22.26 would’ve ranked 15th last year, and the high end of 25.76 would’ve been 11th. This range seems to match up with how the Hoos are ranked to start the season, with most placing them in the high teens (the pre-season AP poll has them at #11). It’s worth noting that the Hoos had an adjusted efficiency margin of 24.29 points in 2017, and finished 12th in Kenpom’s rankings. In many ways, the upcoming season will be very similar to that one. That team was moving on from the graduations of Malcolm Brogdon and Anthony Gill and featured many new faces, namely the core of the team that just won the national championship. The 2017 Hoos were second in adjusted defensive efficiency, but just 50th for offense, something that the 2019-20 Hoos will likely mirror at times. 

All in all, while we should definitely not be expecting another national championship, there is enough evidence to suggest that the Wahoos will still be a very good basketball team in 2019. Tony Bennett has been so successful as a coach that even a team expected to be among the 15 best in the country will be a relative letdown compared to last year. I’m very confident that any Virginia fan would’ve taken a team ranked in the low teens in a heartbeat just a decade ago. This is truly the golden age of Virginia basketball (and Virginia sports in general), and all fans should be ecstatic about the future of the team under the guidance of Coach Bennett and his staff. The one area of concern that many do have for the season is 3-point shooting. Guy, Jerome, and Hunter all shot 40% or better from deep last year, and were reliable in many situations to just create offense out of nothing. I’m not sure anyone on the roster fits that role this year, barring a breakout from highly touted first-year Casey Morsell or transfer Tomas Woldetensae. Jay Huff is a career 42% shooter from downtown, but on only 38 attempts. Can he keep up that pace on a full season? Tomas Woldetensae shot 47.6% from 3 last year in JUCO ball, but will that transfer to the ACC? There are definitely a lot of offensive question marks for the Hoos in 2019, but the one guarantee is that the defense will once again be excellent. Mamadi Diakite led the ACC in block percentage last year, and Braxton Key and Jay Huff were not far behind. Kihei Clark will always be a pest for opposing point guards, and Casey Morsell comes in with a reputation of being a lockdown defender. At this point, it’s best to put our trust in Coach Bennett and hope that the Hoos will once again be competing deep into March.

Leave a Reply