Lies, Damn Lies, And Statistics

Photo by jmvnoos from Flickr

The ability to score is basketball’s most celebrated and widely discussed skill. At a very base, essential level, a player’s gift for creating points seems to be the default framework from which we judge their relative value. This can even be seen at your local gym, during any given pick-up game. It’s rare that a great interior defender or hustle guy is lauded for his contributions. Instead, we marvel at those who never miss three pointers, or have an uncanny ability to get to and finish difficult shots at the rim.

This seemingly innate obsession with scoring led to an inefficient distribution of wealth amongst NBA teams and players. Volume scoring, was valued at a disproportionate rate. While defense and rebounding were being generally overlooked (at least when it came to player compensation).

What sprung from this utter miscalculation and misinterpretation was an intellectually stimulating, numbers driven approach to evaluating a basketball player’s contribution to winning games. Bill James’ Win Shares, John Hollinger’s PER, advanced +/- stats, and Synergy’s point per possession stats combined with an incredible video library, all serve as invaluable tools for better understanding the complex, multi variable puzzle that is basketball. We have more and more data to either back up or contradict what we witness with our own eyes.

While statistics and numbers are an indespensable tool when evaluating the impetus for team and player success, they also carry with them glaring holes and inconsistencies. Win Shares would have you believe you could win a championship with 5 Ben Wallaces (or even more absurdly that Dennis Rodman was more valuable to the Bulls than Jordan), +/- requires a very large sample size and isn’t entirely reliable, and we are still in the infantile stages of advanced individual defensive metrics. Furthermore, stats aren’t yet sophisticated enough to fully account for the effect a players’ context has on his performance, or how his performance changes the context (positively or negatively) for his teammates. Stats can’t always provide the reason for why certain things occurred, only that they occurred (this is what makes Synergy’s combination of stats and video so incredibly awesome).

As one would expect, volume scorers have been the greatest “victim” of the analytics movement. The numbers have allowed us to see that high point totals are not necessarily indicative of a valuable player. However, it does feel as though, for certain players, the pendulum has swung too far the other way. Where as before we were overvaluing and overpaying elite volume scorers, we are currently ranking any slightly above average defender ahead of a talented bucket getter.

One of the most highly discussed, and fiercely debated topics (at least last summer anyway) is whether or not Carmelo Anthony should be considered an elite player. There is no doubt that throughout his career he has lacked the mental focus and commitment on the defensive side of the ball. No one is disputing this. Still there remains significant questions about whether the man who many call “the most gifted scorer in the league” is actually helping his team on the offensive side of the ball.

Even the most basic of statistics seem to indicate that Melo is little more than average when it comes to scoring at an efficient rate. His TS% has never been particularly high, and most of his career he’s taken far more threes than his conversion rate would dictate he should (though he did shoot 42% from deep after joining the Knicks last season). Furthermore, his assist rate would indicate that Melo lacks the ability to create scoring opportunities for his teammates on a regular basis. Thus Anthony seems to fall into the dreaded Iverson category: a phenomenally talented individual who’s popularity was greater than his actual on court contributions. This is where the advanced metrics are failing to capture the entire picture.

While Carmelo may not be directly assisting on teammates baskets, the defensive attention he demands creates rotations issues for the opposition, which in turn allows others to flourish. Many teams have adopted the strategy (including the Lakers in the 2009 WCF) of pre-rotating a big man to Melo’s side when he catches the ball. In 2010, the Jazz where essentially using a 3 man zone to deter Melo from going to the rim. Nate Silver, in a very interesting post over at the New York Times Off The Dribble Blog, pointed out that many players gain a significant boost in TS% when on the floor with Anthony. Watching the Nuggets play over the past couple years would seem to confirm this theory. While Aaron Afflalo is without a doubt a talented shooter, he also benefited greatly from more than a few wide open three-point looks created by the extra attention dedicated to Melo. And Nene was able to obtain such startling levels of efficiency in part due to Anthony’s willingness to carry the brunt of the offensive load. Is Carmelo an elite player overall? Not really. Until he commits himself on defense and shows a greater willingness to pass, it’s impossible to put him the class of Lebron and Wade. Still, Anthony is an elite offensive talent who could serve as the go to scorer on a championship contender.

The debate over Carmelo Anthony highlights an important flaw in the current basketball discourse. Mark Twain once famously said, “ There are lies, damn lies, and statistics.” This quote is often misconstrued. It’s not that that Twain hated numbers or had an evil math teacher. Twain was instead referring to the persuasive power that lies within a statistic. Stats and data are often treated as dogma, hailed as infallible truth. You can be characterized as a fool for going against what is declared to be “indisputable evidence”. The fact is that statistics can and will lie to anyone. Furthermore, stats have biases. The designer of the study often brings his own conceptions and prejudices about what should be considered valuable, yet the conclusions are still treated as pure, scientific analysis. The phrase “but the numbers say…” should never be the end-all-be-all to anything as complex and multi-faceted as individual player analysis in basketball. There’s a lot to be gained from statistics. They aren’t in any way ruining the “magic” of sports. But the blind acceptance of statistics is just as obnoxious and dangerous as the “he never played” qualifier.