Profit from NBA stats? As easy as drawing water from a stone
Thu, Feb 22, 2007
By Tim Roberts
I’ll admit that I’m a sports nerd. I went straight from reading Tintin graphic novels to reading Bill James’ Baseball Abstract as a kid.
I’m guessing that a lot of people reading this are the same way. Don’t worry, I won’t blow your cover (have fun doing research for your fantasy baseball draft this weekend).
Anyways, earlier this winter I heard spoken what I considered to be near-blasphemy.
“To be totally honest, NBA statistics are borderline meaningless,” Covers Expert Ted Sevransky told me in January. “I use stats less with the NBA than I do with any other sport. Nothing’s worked for me, not long-term anyway, and I haven’t found anyone who has used an NBA stat successfully over a long term.”
It was a case of deja entendu on Wednesday when I spoke with Covers Expert Steve Merril.
“The NBA’s the only sport I handicap that I almost don’t use stats for,” Merril says. “From top to bottom, in most stats there’s hardly a difference between the teams.”
“One thing I’ve tried to find for years is a better use of the NBA box score. Football box scores that include drive charts really tell the story of a game, but there’s no comparable use for an NBA box score to get the same level of meaning out of it.”
It’s not surprising, given the nature of the sport.
Baseball is a statistician’s haven. Every play starts with an action (the pitch), followed by a reaction (the batter’s decision to swing or not swing) and a quantifiable result. The repetition of the pattern gives number-crunchers a massive sample from which meaningful percentages arise.
Football isn’t too far removed from that. Instead of measuring three categories (action, reaction and result) as occurs in baseball, football stats can be broken into five general groups (run offense, pass offense, run defense, pass defense and special teams), all of which have meaning to analysts and bettors.
Then there’s basketball: dictated by rhythm, flow and intermittent chaos. Points are scored in so many fashions and forms it’s nearly impossible to devise a formula to predict future outcomes based on past results.
"When a basket goes in, you have no idea how to apportion the credit," Dan Rosenbaum, economics professor and consultant for the Cleveland Cavaliers told the Bay Area's Mercury News. "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."
That means as hard as it is for NBA teams to assess the sport with numbers, it's even harder on NBA bettors to devise a winning formula. After all, the teams have nothing but time and money to throw at the problem and they still can't crack the code (if such a code even exists).
“It really varies game-by-game in the NBA more so than in other sports,” Merril says. “The current form of a team is key and over the years I’ve noticed that teams tend to run good and bad in five or six-game runs. So I try to catch teams during good spells and avoid them on the downswing.”
It's not just bettors who are at a loss. NBA teams are trying to find that winning equation, with different levels of success.
Mark Emmons of the Mercury News wrote about teams hiring mathematicians among others to formulate a “magic number”.
“We take this very seriously,” Pete D’Allessandro, director of basketball operations for the Golden State Warriors, told Emmons. “If you’re not doing this, then you’re not keeping up.”
Emmons briefly described the work of economics professor David Berri, co-author of The Wages of Wins based on his "wins produced" statistic, as well as various other formulae (e.g. ESPN.com’s John Hollinger and his "player efficiency rating," 82games.com’s Roland Beech and his "on-court/off-the-court," the Dallas Mavericks’ "WINVAL" method) before noting that as they get hired by teams, statisticians' work leaves the public realm.
D’Alessandro has overheard Beech and other stat gurus debating their various methods and suggests that we’re still quite a ways from achieving statistical nirvana.
“Listening to them banter back and forth, I’m always learning something,” he told Emmons. “But I’ve also noticed that neither of them are in agreement on anything. Obviously someone is not right.”
All of which leads us back to the old-fashioned methods for basketball handicapping: reading local newspapers and blogs to get the inside scoop on a team and watching as many games as you can fit into your days.
That doesn’t mean that people won’t still kill themselves trying to establish the basketball equivalent of MLB’s "win shares."
Of course, people died trying to locate the Fountain of Youth, too.