"Moneyball" Statistics In Use, But Not Yet Perfect

Mark Emmons of the Mercury News wrote an article about basketball's new "Moneyball" statistics, with a heavy emphasis on David Berri's Wages of Wins.

The article quotes a Warriors official saying he can't imagine there's a team left in the NBA that isn't following the new breed of statistics.

"The Wages of Wins," published by Stanford University Press, argues the best way to judge players is through a relatively simple formula that uses statistics found in box scores and results in a Wins Produced number.

Cuban wrote that the premise is "ridiculous" because "it completely discounts the value of coaching."

Berri doesn't seem bothered by criticism. He says he has no interest in working for an NBA team. And he understands the scorn that would be heaped upon a general manager if he, say, traded Denver star Carmelo Anthony for New York reserve David Lee.

"They would put a show on ESPN Classic about how it was one of the dumbest deals of all-time,'' he said.

Still, Berri ranked Lee as the NBA's sixth-most productive player at this season's halfway point.


One quandary faced by teams, added the Warriors' D'Alessandro, is the lack of uniformity among the numbers guys. D'Alessandro is friends with Beech and Oliver, and sometimes has gone to lunch with them.

"Listening to them banter back and forth, I'm always learning something,'' he said. "But I've also noticed that neither of them are in agreement on anything. Obviously someone is not right.''

UPDATE: Dean Oliver sets the record straight in a comment:

Oh sure, quote the part of the story that is wrong. Pete didn't say that Roland and I always disagree. He called me today to apologize for that somehow getting quoted that way. It's the media and sometimes things get out there wrong. Roland and I both shook our heads in light-hearted confusion at seeing it.

In general, the stat community is going the same direction. The basics of what we do are pretty common -- with Dan Rosenbaum, Justin Kubatko, Kevin Pelton, and I putting a paper out in JQAS (Journal of Quantitative Analysis of Sports) summarizing the commonalities. With the difficulty of analyzing interactions, there will be continuing disagreements over some players, but a general rule of thumb is: Dean is right... ;)