NBA Gets Warm & Fuzzy With Math

NBA Gets Warm & Fuzzy With Math

<div class="quote_poster">Quote:</div><div class="quote_post">The cheering of Bay Area fans over the Warriors trade that shipped Mike Dunleavy and Troy Murphy -- along with their hefty contracts and headbands -- to Indiana last month continues to echo. The deal, which brought Al Harrington to Oakland, was wildly popular.

It shouldn't have been. Or so believes David Berri.

``Harrington at power forward is a disaster,'' Berri said. He called Harrington one of the five least productive regulars in the NBA last season, a player who will not improve the Warriors, and ``won't help his next team, either.''

Berri points to the Warriors' 6-8 record since Harrington and Stephen Jackson joined the lineup and adds: ``That's exactly what you would expect.''

And who is Berri? A guy whose sports career ended in junior high, Berri is a Cal State-Bakersfield economics professor and co-author of the book ``The Wages of Wins.''

He is part of the statistical analysis trend that is gaining an NBA foothold. Think ``Moneyball'' on the hardwood. Just as the numbers-centric view of evaluating players championed by the A's Billy Beane has revolutionized baseball, more NBA franchises are utilizing sophisticated algorithms to better understand their sport's X's and O's.

Armed with spreadsheets and dizzying formulas with names such as WINVAL, ProductivityValue and Player Efficiency Rating, some very smart people are trying to measure an individual's contributions in a complex team game.

``We absolutely take this very seriously,'' said Pete D'Alessandro, the Warriors' director of basketball operations. ``If you're not doing this, you're probably not keeping up. I can't imagine a team out there that at least isn't paying attention to it.''

Some franchises are adding front-office evaluators who know far more about snapping pencils than breaking ankles. The Houston Rockets' general manager in waiting, Daryl Morey, is a graduate of MIT's Sloan School of Management who has no coaching, playing or scouting experience.

The stat geeks are checking into the game. And the conclusions of some challenge conventional wisdom. Exhibit A: Berri's opinions about Harrington.

Quantifying game

There's a long history of stat-crunching in baseball, where sabermetrics -- derived from the acronym SABR, which stands for the Society for American Baseball Research -- is something of a religion. Quantitative analysis of baseball essentially involves computing the results of one-on-one confrontations between pitchers and hitters.

But basketball is more complicated on account of its interdependency among five players.

``When a basket goes in, you have no idea how to apportion the credit,'' said Dan Rosenbaum, an economics professor at the University of North Carolina-Greensboro and a consultant for the Cleveland Cavaliers. ``Some should go to the guy who scored, some to the guy with the assist, some to the player standing at the three-point line who drew defenders to him and created open space. It's even harder to assess who should get credit on defense. That's why evaluations have been subjective.''

Now, other -- and provocative -- tools are emerging. Traditional box scores can be inadequate in judging a player's performance, so statisticians are devising new measures. They comb data in search of a new sport buzz word: efficiencies.

Some, such as ESPN.com writer John Hollinger's Player Efficiency Rating, combine a range of stats to create one number they believe best captures a player's effectiveness. Others borrow from hockey's plus-minus method. So, if a team scores two points more than the opponent while a player is on the court, he would have a plus-2 rating.

But that basic concept is only the starting point for complex equations that could make a fan's head hurt.

Roland Beech works out of his home office in Aptos, where he endlessly watches games -- often in slow motion -- and charts stats. He has created an adjusted plus-minus measurement he calls ``on-court/off-the-court.''

``If you play for the Memphis Grizzlies, your plus-minus is likely to be negative no matter who you are,'' said Beech, founder of the Web site 82games.com. ``But does the team play better with you or without you? With this, you can get a sense of if a guy is actually improving things.''

Such systems are gaining credibility. A few years ago, Beech was one of several people who were doing analysis on the Internet fringes. ``Now, most of my friends from back then work for teams,'' he said. So does Beech, who declined to identify the franchise. Dean Oliver, who played at Cal Tech, has a doctorate degree and worked as an environmental engineer, was hired last year as the Denver Nuggets' director of quantitative analysis.

The Dallas Mavericks also were early adapters. The Mavs reportedly use WINVAL, a modified plus-minus method devised by Jeff Sagarin and Wayne Winston, one of owner Mark Cuban's college professors.

``It's critical,'' Cuban wrote in an e-mail. ``But of course, having seen the other systems, I think ours is the only one that's worthwhile. The primary thing to note is that if you are using publicly available NBA stats as your only source of data, it won't be of any value.''

As people like Beech, Rosenbaum and Oliver have joined teams, their best work is no longer public.

Teams are secretive, too. Warriors executives Chris Mullin and D'Alessandro both declined to speak in specifics about what their team does. But Mullin puts validity in the numbers . . . up to a point.

``I heard there's a book out where they make the case for the most overrated player: Allen Iverson,'' Mullin said. ``Well, Iverson is pretty good. So when I consider statistical formulas, something like that is a red flag.''

That would be Berri's book.</div>

Source