What should we expect from Harvard this year?

Last year’s season was a mixed bag for the Crimson. The team emerged from spotty non-conference play to dominate the Ancient Eight, going 12-2 and sharing the conference title with the Quakers. Of course, Penn would go on to defeat Harvard in the conference tournament and earn the most coveted prize: a trip to March Madness.

The Quakers undoubtedly benefited from playing the conference tourney on their home floor, a built-in advantage that executive director Robin Harris has decided is worth the trade-off of hosting the Ivy League Tournament at the largest and most historic venue that the conference has to offer. However, while Crimson fans might be apt to cry foul, there is reasonable evidence that Penn was indeed the best team in the Ivy League.

Consider the following rankings of team abilities and homecourt advantages. These results are from a NCAA Basketball Model that I built using historical game scores and game sites. The model has been shown to have reasonable predictive power, such as when it scored in the top 26 percent of entries in the Google Cloud & NCAA March Madness Machine Learning Competition last year. It’s not perfect, but it does a decent job of ranking teams.

The above graph shows the model’s estimates of team abilities after training on the data from last season. The estimates are centered at 0, which denotes the ability of an “average” Division I team. For example, if Brown played an average D1 school at a neutral site, the model would expect them to lose by about five points.

We can see that Penn and Harvard are the only above-average teams in the conference, with Princeton and Yale following closely behind. What some might find surprising is that Penn is rated as a better team than Harvard by nearly four points.

The first chart displayed context-neutral team abilities. This second chart shows the model’s estimates of how much better (or worse, in Brown’s case) teams play at home. Harvard and Cornell get the biggest boosts, adding over four points to their expected score differentials in their home gyms.

The model’s results would suggest that Penn didn’t get to play Kansas in the first round of March Madness simply because the Palestra crowd willed them there; they were just the better team. Playing the conference tournament championship at home added a point-and-a-half to Penn’s expected win margin over Harvard, which was already hovering around four, leading to a total win margin of around 5.5. Not all too different from the 68-65 score that was observed.

Penn played great basketball last season, and it shouldn’t be too surprising that the model thinks they are a good team. To me, the real story is that Harvard wasn’t better.

There has been lots of talk about Harvard’s heralded 2020 class. This group includes the reigning Ivy League Player of the Year, Seth Towns, as well as fellow First-Team-All-Ivy center, Chris Lewis. Both of these players had multiple offers from big-name programs, and were heavily recruited out of high school. Add to them the seven-footer Robert Baker, the elite defender Justin Bassey, freshman-year-scoring phenom Bryce Aiken, and last year’s biggest breakout player, Christian Juzang, and you’ve got a group that ESPN once deemed the 10th-best recruiting class in the nation. Unheard-of talent for the Ivy League.

The 2020 group has been the focus of coach Tommy Amaker’s rotations the past two years, and their talents haven’t gone unnoticed. Rupp Arena witnessed Seth Towns drop 25 points on their Wildcats last year. Lewis dropped a 25-spot on Penn in a crucial game at Lavietes Pavilion, looking unguardable in the paint. There are times when the athleticism of the now-junior class can overwhelm their conference opponents.

Yet, even with their unique abilities, the class of 2020 has been somewhat of a disappointment at times. Last season began with an embarrassing overtime win against MIT, in which the D3 squad featured a more cohesive offense and even a larger crowd. Armed with a fraction of Harvard’s recruiting talent, Penn, Yale, and Princeton regularly post superior offensive efficiencies. As Harvard showcases its staunch defense and mesmerizing individual performances from Aiken, Towns, and Lewis, the question remains puzzling: Why can’t this team score?

Perhaps this season will be the year that Harvard elevates its performance to the next level. Everyone who contributed to last year’s success is returning, and this year’s freshmen are nearly as impressive as the 2020 squad. Noah Kirkwood, who was electric during Harvard’s Crimson Madness scrimmage event, has the makings of a top Ivy scorer. Spencer Freedman shunned USC to come play for Harvard, and he features a pass-first point guard game that could be just what Harvard needs to get its offense rolling. The 6’8 forward Mason Forbes also seems poised to contribute.

Then again, Harvard isn’t the only team in the conference with athletic recruits anymore. Desmond Cambridge cemented himself as one of the best offensive players in the league during his freshman year. Princeton is welcoming Jaelin Llewellyn to the fold, who is easily the preseason favorite to win Rookie of the Year.

Seth Towns and Bryce Aiken both sat out during Crimson Madness, and neither of them suited up for a secret scrimmage against BC on Oct. 19. If these two aren’t healthy in time for the season, it could be a very different outlook for Harvard.