Friday, November 12, 2010
Picking the NBA's best rebounder so far
By Henry Abbott
Gary Dineen/NBAE via Getty Images
By old metrics, Joakim Noah has been the best. But why use old metrics?
Justin Kubatko, founder of Basketball-Reference, consultant to the Blazers, and undefeated champion in his two appearances in the TrueHoop Stat Geek Smackdown, is among the most respected names in NBA stat geekery. He's also, now, writing for the New York Times.
He kicks off his season in advanced stats by holding our hand a bit through some of the tools he uses to look at the game.
He explains that basketball trails baseball, generally in the state of analysis, but also says that "progress on the basketball side has been steady" and "much has been accomplished."
Then he goes into some of the ways simple tweaks to analysis can give us not just different, but better ways to assess the game. I'm with him every step of the way.
For example, he addresses total rebound percentage, which he says is "an estimate of the percentage of available rebounds a player grabbed while he was on the floor. It is an improvement over rebounds per game because it takes into account opportunities, which are influenced by the pace that a team plays and the number of missed shots that a team forces."
Once you know that number exists, it almost seems silly to obsess over rebounds per game, right? I mean, you don't even know how many minutes each person played, let alone how many rebound opportunities they had. If you're trying to assess who is best at getting the ball, you want to know how many chances they got to try, right?
To use a baseball analogy, the way we used to measure this is like saying an outfielder caught seven balls in a game, and is therefore better than a guy who caught six. How many balls were hit to his area? How many innings did he play? What? You don't know?
Or, worse yet, you do know, but decide to ignore all that?
Kubatko is very measured, but you can feel it, just a little, the urge to toss aside some of the sillier older metrics. "With the help of statistics like these," says Kubatko, "fans can get a better, more complete view of the players and the teams than they would with a cursory look at the traditional box score."
So, hats off to The New York Times for egging on the NBA's statistical revolution.
And as for all those suckers who still believe in things like rebounds per game, HA! We scoff at you for living in the dark ages.
But ... Wait a second.
I linked to the online version, but I read this on newsprint, where Kubatko's article appears on page B15, less than an inch away from page B14's "NBA Leaders" listings. Papers have been carrying forever. Four sets of top ten lists, three out of the four expressed the old way. The dumb way: Per game.
And lookie there, we have rebounds per game ... the very thing Kubatko just explained to us is lacking.
That old school list (which is also, literally, old -- it's through November 10) tells us that the league's best rebounder, by far, is Joakim Noah at 15 per game. Some others in the list: Luis Scola (fourth), David Lee (fifth), Pau Gasol (seventh), Lamar Odom (eighth) and Kevin Garnett (tenth).
On Kubatko's excellent Basketball-Reference, I looked up the statistic he prefers, the league leaders in total rebound percentage. Things are pretty darned different. By total rebound percentage, Noah is fourth, and Reggie Evans is the guy well ahead of the pack, followed by Kevin Love and Marcus Camby.
And those other guys I mentioned from the rebounds per game list? None is even in the NBA's top 15 by the better metric, and some of them are off in the distance somewhere -- Scola's 27th for instance, and Gasol is 34th.
Meanwhile, the paper's list ignores people like Samuel Dalembert, Marcin Gortat, Ben Wallace, Derrick Favors and Andris Biedrins, who are all in the top ten by the more useful measure.
Am I the only one getting mixed messages from my morning paper?
I realize I'm ruthlessly picking on The New York Times here -- they're doing what everybody does with these lists, and they're about as smart as anybody with stats in general. But if stat geekery as a movement has villains, these lists are surely among them. People have long been hypnotized -- and owners have long overpaid -- for players ranked highly here. Curing that inefficiency drives the field.
Serving up both kinds of content side-by-side shows that things are changing, but there's still plenty of room to grow.