March 12, 2014

ZAKOUR | Analytics’ Place in Basketball?

Print More

By John Zakour

A recent Deadspin piece entitled “Two Days At Sloan: How Sports Analytics Got Lost in The Fog” got me thinking. The authors claim basketball analytics are lost in the fog. With that, I resoundingly agree. But I also believe they are being overvalued by fans. And by fans I should say fans in my demographic. The generation that grew up with Moneyball and rolls its eyes when announcers call OPS an “advanced stat.” Amongst the fog of analytics, it is easy to lose the forest through the trees.

I love baseball. I cannot wait for the season to start and for my Mets to disappoint me. I love sabermetrics and advanced stats. For quick and dirty analysis, WAR is a godsend. I believe Mike Trout was robbed of two MVPs. So yeah, I am one of those.

The part that makes sabermetrics worthwhile is that baseball is essentially a one on one contest of pitcher and hitter. Correct for a few variables (park effects and defense), and stats can tell you basically how effective a hitter or pitcher is pretty definitively.

If you go to FanGraphs and sort by wOBA or wRC+, you have a list of the best hitters and position players in the league. Same for pitchers, even though it is still muddled by the effect of variable innings. However, defense still remains murky. It is tough to gauge because it is affected by positioning and random bounces, and it is hard to quantify things like range and a good first jump. But you can still get a decent picture if you have over a season’s worth of data.

So when the advanced analytics movement comes to basketball and, to a lesser extent, hockey and soccer, forgive me if I am not the first in line waving the analytics flag. If that seems incongruent, then let me explain.

I am not sure there ever will be a holy grail of basketball analytics. Basketball is a complex sport, with a lot more movement and noise than baseball. Defense is extremely hard to quantify and governed by so many moving parts. Even offense is pretty complicated in terms of plays, fakes and movements. A significant part of deciding basketball games is just where the ball bounces after a missed shot, sometimes pretty randomly.

That is not to say that advanced basketball metrics are voodoo or worthless. I just do not think advanced metrics in basketball will launch any Moneyball-eque revolutions. Basketball general managers and CEOs are already starting to look more like well educated guys who are just smart and savvy businessmen.

A smart scout that knows a few metrics is just as valuable as the best model basketball can produce at the moment (think of Grantland’s great Zach Lowe). Just because there is finally being research done in sports does not mean it is the be all and end all. In the world of professional basketball, the margin between teams are growing smaller. Every team knows what they are doing to an extent. I am not sure analytics give any one team anything more than a minute edge, compared to the other facets of running a team.

The margins in the NBA are small. Every little bit counts. But analytics might not be that important in the long run. I would much rather put my time and resources into developing my draft picks with the best coaching staff I can assemble, rather than hiring everyone who came off the assembly line from MIT. Every front office is different. An old school one, like the one Donnie Walsh and Larry Bird built the Pacers into the juggernaut with, can have success.

The Houston Rockets are, as far as I know, the most analytically inclined front office in the NBA. At least, they certainly make it seem like they are. Their general manager, Daryl Morey, graduated with degrees from Northwestern for computer science and a MBA from MIT. And while everything you read on Morey will make him seem like the Billy Beane of basketball, I am not so convinced. These articles include Sports Illustrated’s “Morey Takes a More Analytical View Than Other executives,” or the New York Times’ “How Do You Know This Game Isn’t Rocket Science?”

And while the media tries to sell you on how bold and different he is, I do not really believe it. Just look at the current Rockets team Morey has assembled. The Rockets are contending in a brutal Western conference, so this is not meant to discredit Morey at all. I just think Morey is a GM like any other good GM, whose edge happens to be analytics. His big moves have been clearing cap space and stockpiling assets by trading guys like Kyle Lowry, who is now having an all-star caliber season in Toronto, to go after James Harden and Dwight Howard for max money. Not exactly revolutionary stuff.

Just because his edge is in analytics, as opposed to scouting or coaching, does not mean Morey is inherently better, or even that his way is the way of the future. We should not glorify analytics at the risk of eschewing coaching and traditional scouting. All of these go into running a team.

But sometimes, sadly, analytics fall on deaf ears. Even worse than overvaluing analytics is ignoring them totally. Analytics should be used situationally and to support and supplement your other departments. All the stats and percentages tell me a horrible offensive play call at the end game should lead you to isolate your best player and try to have him score.

Yet time and time again I see Kobe or Carmelo Anthony get the ball and watch his teammates clear out. And a lot more often than not, it does not work. The only coach who does not do this with regularity and who has the stars to get away with it is Gregg Popovich, widely regarded as the best coach in the NBA.

This should be the strength of analytics in sports, getting people to rethink the ways they do things. Sometimes we do things just because they have always been that way. And that is not good enough.