The first blockbuster trade of the 2014-15 NBA season came late Thursday, when the Boston Celtics finally shipped Rajon Rondo, their perpetually on-the-trading-block star, to the Dallas Mavericks for a handful of players, picks and a trade exception.
Others have broken down the pros and cons of the trade, but the deal also serves to spark discussion about Rondo himself, one of the most polarizing players in the league. In early November, when Grantland’s Kirk Goldsberry profiled the enigmatic point guard, he touched on the conundrum advanced stats face in measuring his performance. “Rondo’s value is difficult to quantify, in part because he doesn’t fit into our established taxonomy of NBA superstars,” Goldsberry wrote. “He amplifies the goodness around him, but he can’t create it.”
The statistical arc of Rondo’s career bears this out. According to metrics ranging from Player Efficiency Rating (PER) to Win Shares and Statistical Plus/Minus (SPM), Rondo’s best seasons came in 2009 and 2010, when he was surrounded by three future Hall of Famers in Kevin Garnett, Paul Pierce and Ray Allen. In Goldsberry’s terms, Rondo amplified their goodness — or at least, he helped organize it — and his numbers received a sizable boost as well. In 2009, he ranked as the NBA’s 14th-best player by Win Shares and its 12th-best by Value Over Replacement Player.
But since Rondo tore his ACL midway through the 2012-13 season and the Celtics broke up their “Big Three” for good, Rondo’s output has taken a nosedive. Stats such as a .461 true shooting percentage (TS%) might have been excused while returning from injury a year ago, but Rondo has continued to shoot the ball poorly (.422 TS%), turn it over frequently (on nearly 26 percent of his possessions) and score infrequently (he’s averaging 9.4 points per 36 minutes) this season. Abysmal numbers like those would typically warrant a benching, not a win-now return in a blockbuster trade.
Then again, Rondo has never been a typical player. Along with the bad shooting and nonexistent scoring, he also provides his team with one of the game’s most prolific passers, its best rebounding point guard, a highly skilled thief and an elite defender. In other words, there’s precious little middle ground with Rondo; pick an area of the game, and he’s either one of the best or one of the worst in it.
So, how do we weigh the positives of Rondo’s game against the negatives? The popular single-number metrics are all over the place. PER thinks he’s average so far this season; Win Shares per 48 minutes considers him substantially below average. Naturally, Wins Produced likes him for his rebounding prowess (and doesn’t care about his microscopic scoring output), while Box Plus/Minus (an SPM variant) pegs him as above-average but completely because of defense (he’s quite negative offensively). Box Plus/Minus jibes with what ESPN’s Real Plus/Minus (RPM) lists for Rondo this season, but last year RPM considered him good offensively and bad defensively (go figure).
Dallas owner Mark Cuban has always had the Mavericks at the forefront of analytics among NBA teams, so it’s possible he’s found his own way to cut through the fog and divine Rondo’s true worth. But for the rest of us, the debate over Rondo is one of how best to measure a unique player. Advanced as they are, our best statistical formulas are trained on trends that emerge over the entire sample of NBA players. Rondo might just be the extreme case that causes our math to break down.