Thursday, September 1, 2011

Shorting Harvard

The Crimson is a Top 50 team. It's a lock to grab the Ivy League title. It'll be a tournament sleeper.

But locks don't have gaping flaws. This Harvard team might.

Offensively, the Crimson is already one of the best teams the Ivies seen over the past 15 years, posting top six Adjusted Offensive Ratings for both the whole season and during the conference slate. Losing no one, adding a stellar recruiting class and getting a full year of Kyle Casey at 100 percent should at the very least guarantee the same, if not better, production during the 2011-2012 campaign.

Harvard hasn't been especially proficient at generating stops, though. It was a pair of final possession buckets which sunk the Crimson at Yale and then again in New Haven against Princeton in the Ivy playoff. Then, there was the victory at The Palestra, in which Harvard failed to stop Penn's last gasp at the end of regulation and the first overtime (well, maybe they did actually stop the latter) and nearly lost the game because of it.

The best Harvard defensive team of the past 15 years was the 2001-2002 edition, followed by the 2000-2001 team. The 2010 and 2011 Crimson squads are the third and fifth stingiest defenses Cambridge has seen during that span (22nd and 31st out of 120 teams overall), but that doesn't change the fact that last year's Harvard squad was the fourth worst defensive team to win at least a share of the Ivy title.

If that were the only concern, however, there wouldn't be much to worry about. After all, the 2009-2010 Big Red wasn't even three points per 100 possessions better on the defensive end and it cruised to an Ivy title behind the one of the best offenses the league has seen in the AI era. As mentioned above, given that the 2010-2011 Crimson finished sixth on that same list (second in conference-only play) and should close the gap even further this year, an average defense probably wouldn't hold Harvard back.

The issues arise when parsing the data a little further.

The data in the table above is adjusted for opponents' quality, meaning that the difference is not merely that the offenses are better in Ivy play than during the non-conference slate - not that such reasoning would make much sense anyway.

One thing to consider is that, on average, Ivy defenses are a point per 100 possessions worse during league play than during the non-conference slate. Looking through the aggregate data, the reasons for this include a slight rise in free throws per field goal attempt during league play and a decline in turnovers, which helps boost offensive efficiency.

Amaker's teams have been a full eight points per 100 possessions worse on the defensive end in Ivy play, continuing a trend of under-performance by Crimson squads that dates back to the beginning of the sample (Frank Sullivan's teams were four points per 100 possessions worse on defense in conference vs. non-conference action). Last year, Harvard made up for the defensive collapse by rising from the 15th best non-conference offense of the past 15 years (102.2) to the second-best Ivy offense during that span (112.5), behind only 1997-1998 Princeton.

Taking a further look at the 2010-2011 campaign, in its first 16 Division I games, the Crimson held opponents to one of their 10 worst offensive performances of the year seven different times. Over its final 12 games, Harvard held just Columbia to a bottom 10 offensive performance, while 10 of the other 11 teams had above average offensive outputs against the Crimson.

Examining Dean Oliver's four factors, the culprit was the one you would least expect. Harvard allowed nearly the exact same eFG% over the first 16 contests vs. the final 12. Rebounding rates and free throws allowed were slightly worse during the final 12 but hardly significantly. The biggest difference was a turnover rate drop from 21.5 percent of possessions to just 17.6. Surprisingly enough, that directional difference has existed during each of the past four years (22.4 percent to 20.8 in 2010; 21.7 to 20.5 in 2009; 20.3 to 18.2 in 2008).

Obviously, the fewer possessions that end in a turnover, the more shots get to the rim either from the field or from the line and the more points an opposing team can score, even if its shooting percentage doesn't improve. Harvard's ability to prevent the latter from happening last year was completely circumvented by opponents simply holding onto the ball well enough to get more shots.

There is no silver bullet for defensive improvement. Focusing more on creating turnovers can leave defenders out of position and can afford the offense easy buckets or more trips to the line via recovery fouls. Focusing more on challenging shots and zone/help defense can make it more difficult to box out, leading to more chances for the opposing team. In such ways, the defensive four factors are all related.

Regardless of the cause, Harvard's league vs. non-league splits on the defensive end are alarming. With a potentially historic offensive arsenal, however, the Crimson might be able to survive another collapse on the defensive end to repeat as Ivy champions.

For Harvard to be labeled a lock, though, that requires a belief that this will finally be the year when the Crimson's solid non-conference defense carries over into the league campaign. That's a decent sized leap of faith for something being sold as a certainty.


  1. how do you see the highly ranked freshman class fitting in?

  2. Amaker has done a great job getting freshmen involved from his first recruiting class in 2008-2009. This might be his best class of his four at Harvard, but it will also be his toughest class to integrate into the lineup, as his returning talent has never been better.

    The Crimson really needed frontcourt depth last year. A lack of confidence in Jeff Georgatos and Andrew Van Nest at times left Harvard playing Laurent Rivard at the 4 - something that just can't happen this year if the Crimson plans to reach the level of success many have charted out for it. Luckily, the freshman class is loaded at the 4/5, with three options there and a fourth (Wesley Saunders) that is probably more suited to be a three, but could also pass as a four.

    It can take some time for freshmen to learn how to defend at a Division I level, so whether the rookies can help there remains to be seen. At least a couple should be good enough to provide an offensive boost over last year's bench, which would bolster the argument that the offense alone might be good enough to overcome any defensive lapses.

    Also, I'd be remiss if I didn't include the usual caveat about freshmen. Harvard's two highest ranked recruits in 2008 (Max Kenyi and Andrew Van Nest) and two of its top three guys in 2010 (Jamie Moore and Ugo Okam) have combined for a high of 47.5% minutes played in a season (Kenyi 2009) and a high offensive rating of 91 (Van Nest 2010), so be careful projecting too much onto the unknowns.