Many of the stickiest challenges in increasing the quantitative understanding of basketball come back to one simple factor: everything is connected to everything else. In contrast to the more neat and tidy world of baseball where a simple interaction (from a conceptual, not physical or mental skill standpoint) between hitter and pitcher is identifiable and possible to isolate, the dynamic, flowing nature of basketball makes breaking finite elements out for study more difficult. Most analysis is still done on this smaller scale, usual at the level of a single possession or play.
For many purposes, this is perfectly satisfactory. The elements which go into the success or failure of a given scoring attempt are reasonable well-understood, if not entirely measured. However, in other areas of analysis the choice of artificial endpoints at the beginning and end of a single unit of play obscures more than it illuminates.
One problem with this extreme focus on the ball itself and the player in possession of it is the contributions of the other nine players on the floor are easy to miss. Even some of the best measures of individual defense include large helpings of luck along with whatever slice of skill is identified. At the same time the value added by playmakers on the floor or play callers from the bench.
Those are issues just within a single unit of possession. A larger problem which is usually intuitively understood but usually analytically ignored is the interrelation between possessions. A team’s offensive possessions are likely not nearly as independant of each other as are at bats in baseball. What a poker player might term the “meta-game” causes the success or failure of one play to influence how the defense reacts to the next, which in turn alters the one after that on down the line.
Further, how smoothly a team switches from offense to defense or vice versa has tremendous impact on that immediately ensuing possession. For as long as there has been basketball on televisions, commentators have correctly argued teams can get themselves more easy opportunities on offense by getting stops on defense. One of the main units of defensive measure, the steal, has come to be known as an offensive stat almost as much as a defensive one given the likelihood of easy points the other way on the sudden change of possession. On a more mundane level, teams get into their own ensuing offense more quickly after rebounding an opponents’ misses than when taking the ball out of the net following a make.
Naturally, the reverse is also true. Bad offensive possessions can compromise a defense before the play even starts. Some coaches consciously instruct their teams to avoid zealous pursuit of offensive rebounds in order to better set their own defenses. Clippers coach Doc Rivers is one such coach. Despite the rebounding prowess of DeAndre Jordan, the Clippers are 27th in the NBA in offensive rebound percentage. In 2013-14 they were 20th. Rivers’s last three Boston teams ranked dead last in in OREB%.
Whether this is a good strategy or not is a bigger question, but it must be acknowledged it is effective for it’s intended purposes. Using SportVU rebounding data, it’s possible to derive an approximation of which teams “chase” the most and fewest offensive rebounds. Through games of March 14, the Clippers had the lowest such “Chase %” in the league, contesting only 35 percent of available misses – as compared to league average of around 43 percent and Minnesota’s league high of 47.3 percent. Comparing this percentage to each teams’ propensity to give up fast break attempts (as defined by Synergy, discounting putback attempts to give as much of an apples-to-apples comparison as possible), teams which attack the offensive glass less are better at getting back:
Similarly, teams which commit fewer live ball turnovers also allow fewer transition opportunities:
It almost goes without saying that forcing the opposition to play half-court rather than fast-break offense is a win for the defense. A quick and dirty regression of these two factors suggests about one extra transition opportunity allowed for every three steals allowed or every five offensive rebounds chased, everything else being equal.
Of course these are hardly the only factors in how imperfect offense can put additional pressure on a defense. Poor transition defense is usually blamed on lack of effort or hustle, which is accurate in a good many cases. But not always. From a usefulness standpoint the determination of whether the question is a team or player not playing smart enough rather than hard enough is crucial to rectifying the problem.
Take Minnesota’s league worst transition defense. The Wolves don’t help themselves with their high chase percentage and their propensity to cough the ball up. But even aside from these issues, the Wolves end a lot of possessions with bad floor balance,such as in this particular egregious example:
Not only does Minnesota have four players flat along the baseline (including Andrew Wiggins who probably has responsibilities to get back as the other wing is taking a corner shot) as the shot goes up, but the fifth player, rookie Zach LaVine decides to attack the offensive glass from the top of the key:
Naturally, this led to a breakaway layup for the Kings and yet more questions about the Wolves defense. But more accurate analysis would rewind the tape just a little bit further and find the bad defense was caused first and foremost by the poor conclusion of an offensive possession, where not only was LaVine’s decision to go to the board a terrible one, the defense was already compromised by virtue of an offensive set which gave four Kings head starts on their defenders once play shifted to the other end of the court. Without looking in the right place, a coach might simply lambaste his teams lack of desire without addressing the underlying dynamics at work and thus never really solve the problem.