It’s an old adage, one so frequently used its been propelled beyond the realm of clichés. Pick any team sport &- you can’t avoid it, or at the very least a closely related variation of it.
Defense wins championships.
Yes, offense is important as well. The stat head in me looks at what “Defense wins championships” insinuates &- that offense, for whatever reason, doesn’t win championships &- and wants to scream. Still, the literal meaning of the aphorism is verifiably true. And while shoddy defense can be compensated with by outstanding offense, some degree of defense is always required for any sport in which defense exists.
How defense is best evaluated depends on the sport &- what is the defense’s goal, and is there a quantitative approach to determining how well the goal is realized? In basketball, the defense’s primary goal is fairly straightforward – prevent scoring. While there is some degree of interplay between defense and offense &- different offensive and defensive strategies affect the types of transitions linking each phase &- the defense’s purpose, narrowly defined, is to keep scoring to a minimum.
What is the best way to measure this? The immediate answer might be points per game. After all, an individual game appears to be a reliable, standardized unit of measurement. If Team A regularly allows fewer points per game than Team B, Team A has a better defense. Right?
Probably &- but the relevant unit isn’t an individual game. It’s an individual possession. Defenses try to prevent points from being scored, but the quantity of points scored in a single game is dependent on more than an individual defense’s level of performance. The number of opportunities to score is also important.
Revisit Teams A and B. Suppose Team A allows 95 points per game, while Team B allowed 100 points per game. It might look like Team A has the better defense &- but what if Team A’s games feature 90 possessions per team while Team B’s games feature 100 possessions per team? Team B has to deal with more opportunities to allow points. Given that context, Team B is defending better at a per-possession level &- on any given possession, Team B is likely to allow fewer points than Team A.
This raw per-possession data can then be shaped into a predictive tool by incorporating adjustments for the factors that demonstrably increase the data’s predictive value: when a game is played, where a game is played and against whom it is played. The final result, expressed in terms of points allowed per 100 possessions, is called adjusted defensive efficiency, and can be found online rather easily &- just Google “Ken Pomeroy.” Naturally, it has a counterpart, adjusted offensive efficiency, which measures the same thing for offenses.
Why is this important? With the right tools, it becomes much easier to determine where a team excels and where it does not. And on a macro level &- just looking at offense and defense instead of their components &- a lot of teams appear to be misjudged.
Take BU men’s basketball, for example. The Terriers’ games have featured more than 70 possessions on average, easily the most of any America East member. As a result, BU ranks just sixth out of nine America East teams in scoring defense.
However, adjusted defensive efficiency places BU second in the conference, slightly behind Maine, and among the top 100 teams in the country. When you look at what the Terriers’ defense has done, this makes a lot of sense. BU has done an outstanding job forcing missed shots and has cleaned the defensive glass better than anyone could have expected back in November. Forcing missed shots and grabbing the rebounds is a pretty good way to prevent points.
But that’s not the real surprise. BU’s defense has actually outstripped the offense. Offense is down all across America East, so the Terriers’ adjusted offensive efficiency still ranks third among conference members, but on a per-possession level the offense is scoring less than last year. When BU’s offense has played well, the results have been memorable &- but that says more about the defense than the offense.
The average NCAA Division-I team scores 100.8 points per 100 possessions, or about a point per possession. Consider one point per possession to be a benchmark. The Terrier offense has exceeded that mark on 11 occasions this season. In ten of those 11 games, the defense has held its opponent below a point per possession, usually by a significant margin. As a result, a quality effort by the Terrier offense has been likely to produce a blowout.
In contrast, while BU’s defense has held opponents below the efficiency midpoint 18 times, eight of those games also saw the Terrier offense score less than a point per possession. As a result, the Terriers’ best defensive efforts aren’t necessarily associated with their best overall performances.
But even if the naked eye may lie about where the Terriers excel, the players themselves know where their work is paying dividends. After BU put up its highest point total of the season in last week’s 93-51 demolition of Binghamton University, senior co-captain Tyler Morris noted, “As long as we expend 100 percent of our energy on the defensive end, the rest will take care of itself.”
Apparently Morris also believes defense wins championships &- and he might be right.