ESPN NBA

This weekend is the annual international confab of NBA stat geeks, the Sloan Sports Analytics Conference at MIT, organized by none other than Houston Rockets GM Daryl Morey. After having a tremendous and enlightening time there a year ago -- there is literally no better collection of NBA smart people anywhere in the world -- I'm thrilled to be heading back.

And I have statistics on the brain.

More and more I hear from people who essentially have the position that newer stats like PER, plus/minus, Win Shares and the like ought to go back where they came from.

They'll point out one exceptional flaw -- the Bucks were briefly seventh in John Hollinger's power rankings, how could we ever trust analytics again! -- as proof that all new stats must be ignored.

When people say things belittling statistics, quietly, inside, my inner counselor gets cranked up: "It's OK. You can admit you don't understand them. I don't understand a lot of them either."

Because there is not a person on the face of the planet who both understands the sophisticated analysis of basketball and thinks it's a fad. It's here to stay, and the sooner you'll embrace that reality, the faster you'll be able to gain useful insight from it. 

Now, read carefully. What I am NOT saying, and really just about nobody is saying, is that new statistics are anywhere close to good enough to replace being a knowledgeable observer of basketball, and watching yourself.

What new statistics are plenty good enough to do, however, is to replace the old statistics. Points and rebounds have always been measured ... why? Because they are the most important things in basketball? No. They have always been measured because ... they are easy to measure.

Defense is important in basketball too. But it has never really been measured, because it's tough to measure.

There are a thousand things that new statistics can suggest, however, which watching the game can explain. For instance, various forms of plus/minus suggest that Andre Iguodala is more effective than is obvious, and that Kevin Durant gives up a ton on the defensive end. Is there something to that? 

Teams are getting answers to all kind of questions. Only in the age of the spreadsheet can coaches know, beyond shooting from the hip: Which point guards turn the ball over the most when you trap the pick and roll? Which free agent power forward is the most efficient offensive rebounder from the weak side? Which players score at a vastly lower rate when going left?

Some of the people I affectionately call "the stat geeks" have real answers to those kinds of questions. And while the coaches of teams with in-house stat geeks and custom databases might not change their routines quickly, don't you hope they spend the time learning about how the best of these statistics can be implemented? Would you really rather your team make decisions based on box-score statistics?

I think basketball analysis in 2009 is like medicine in 1900. It's new, and some of it is crackpot. Yes, there is probably more dogma then their ought to be from the true believers, but also too much skepticism from the old guard. Even more than a century later, you can make a case that some modern medicine is harmful.

But let's look at the big picture. It is 2009. And if you have a major medical problem, you are wise to consult with an array of doctors, who have all kinds of tools, some of them difficult to understand, at their disposal. Warm though you may feel about home bloodletting, we can all agree it was never a smart solution, right? Medical science was not, it turns out, a fad.

Things evolve. Right now, basketball statistics are evolving. There's nothing anybody can do to stop it, and out of all this, in the not-too distant future, will be arriving some essential and lasting new developments in how we see the game. Go ahead, dip your toe in.

(And I promise TrueHoop will try to make that a little easier for you. Hopefully there'll be plenty more sober statistics to come.)