Freakonomics in The Times Magazine: Hoop Data Dreams

Levitt and I have a column in this Sunday’s Times Magazine about the attempt to bring to the sport of basketball the intense statistical analysis that Bill James has made popular throughout baseball.

The column centers on the Boston Celtics, who have just completed the best-ever turnaround in N.B.A. history, winning 66 games this year after winning just 24 last year. The Celtics are one of a handful of N.B.A. teams that have recently become very data-centric (including the Rockets, whose general manager, Daryl Morey, came to Houston from the Celtics), and the Celtics’ stat man is a very colorful young guy named Mike Zarren. Here’s a bit about Zarren from the column:

[He] seems to know every data point about every N.B.A. player, past and present. Garnett calls him Numbers, the Celtics Dancers call him Stats and Paul Pierce, the team’s longtime standout, calls him M.I.T. even though Zarren never went there. He did, however, lead a University of Chicago quiz-bowl team to four national tournament victories and later graduated from Harvard Law. (Disclosure: Steven Levitt taught Zarren while the latter was an economics undergrad at Chicago.) …

Zarren also happens to be the team’s associate counsel, although this would be hard to believe if you came across him at a game, way up in the cheap seats, wearing his green satin Celtics jacket and shouting himself hoarse: ”He pushed! He pushed! … You got ’em, K.G.!” To describe his Celtics fandom as rabid would be a gross understatement. This is his third season on the Celtics’ payroll — he worked two years without pay as a law student and while clerking for a judge — but his family has had season tickets since 1974. He began regularly attending games at age 5, and since moving back to Boston after college, he has missed only five home games. ”Three of those,” he says, ”were due to illness.”

Zarren and Danny Ainge, the Celtics’ general manager, were way too smart to tell us much about their own and opposing players’ tendencies, strengths, and weaknesses. They are, after all, trying to win a championship, and even the lowly Hawks are proving pretty pesky in the first round of playoffs. Nevertheless, I hope you all find the piece as interesting to read as we found it interesting to write.

It was particularly worthwhile talking to Ainge, who was a great three-sport athlete and whose basketball intelligence is very keen. He said one thing during our interview that didn’t make it into the piece because it’s really off-topic, but it’s related to the talent/practice question we’ve written about before (most recently in regards to A-Rod).

When we asked Ainge about the relationship between talent, practice, and athletic excellence, here’s what he had to say:

My experience tells me that people like to do what they’re really good at. And so, in my life and the players I’ve been around, it’s unbelievable how I look out there on the court and I watch Eddie House and Ray Allen shooting jump shots in practice. I’ll look in there and I’ll see Kendrick Perkins and Big Baby lifting weights. It should be the other way around, but it’s been that way all the time.

The guys that are great at getting stronger and have great bodies are in there lifting weights and the shooters are shooting because that’s what they like to do. So that’s what my experience tells me. I was really good as an athlete as a child, so I gravitated toward it because from age five, I was really good, and I knew I was good. And I had older brothers that knew I was good and I got a lot of attention for being good and so that’s what I did.

Note: This will be our last Times Magazine column for a while. We are taking a break until we finish writing SuperFreakonomics. We started out writing one column a month, then switched to every other month. But even that is too much of a load, at least for us, when you’re trying to write a book.

This blog, however, isn’t going anywhere.


Many comments about how the Celtic's two big trade made more of a difference than stat analysis. However, stat analysis may have helped to decide what players to give away (ie how much the Celtics can afford to lose ) for the 2 players, or whether they can fit with the other players. In fairness, though, any trade for Garnett seems like a no-brainer, even without hindsight on how big an impact he would be.

Ray G

The great thing about statistics in this kind of scenario is that even if the stats are not done entirely correctly, looking this closely at each player and the various scenarios of player interaction in differing settings will bring about all kinds of discoveries.

One kind of offense working better in a certain setting, et cetera.

It is also very useful come contract negotiation time.

Buzz Lightyear

Michael's comment about team chemistry is extremely prescient.

Let's look at 3 mid-season trades made in the Western Conference this year:

1) The Mavs trade for Jason Kidd...and go about .500 for the rest of the year, then get *schooled* by the Hornets in the playoffs. Plus they've given up an up-and-coming defensive stud PG (Devin Harris) for an old/slow Kidd and his monster contract.

2) The Suns trade Shawn Marion for Shaquille O'Neal...and go into the tank. They 'right the ship' and look pretty good going into the playoffs...where the Spurs take them apart. Only the Suns completely revamping their offense to center around Boris Diaw extends their first-round series.

3) The Lakers, trade for Pau Gasol...and go something like 17-3 with him in the lineup. They fairly easily sweep Denver in the first round and will probably dispatch the Jazz in 5 or 6 games.

One can argue that Shaq, Kidd and Gasol are all NBA players of comparable statistical worth (Shaq's declining stats are somewhat compensated by the premium placed on true low-post players in the NBA).

Yet we had 3 very different outcomes. In the case of Kidd/Mavs, Kidd's rebounding, passing and fast-break skills didn't complement what the Mavs already had, and his declining defense increase their vulnerability to the lightning-bug Chris Paul-type players.

In the case of Shaq, his size and low-post presence improved their rebounding and interior defense, but losing Marion cost them a *lot* of versatility and compromised the Suns perimeter defense (Shaq, Stoudemire and Nash are all poor defenders).

In the case of the Lakers, Pau Gasol's shooting/passing skills and his basketball IQ were *perfect* fits for the Lakers' triangle offense. Teams having to account for Gasol defensively (unlike hands-of-stone Kwame Brown) means easier shots for the rest of the Lakers...or an endless procession of dunks as seen in Game 1 of the Lakers/Nuggets playoff series.

I'm sure stat geeks can measure these effects, but I'm not sure that the can predict them ahead of time.



I suppose statistics might give you some small edge concerning the game but Basketball (more so than any other US team sport IMO) is the sport most dependent on team chemistry - which is difficult for anyone to measure.


Very interesting read

and Frank, are you serious? Actually, every REAL NBA fan loves the "where amazing happens" slogan. I have not met one single NBA fan who doesn't love that slogan. This slogan is having a huge success and fans are loving it and are coming out with a lot of different "where amazing happens" videos..just check out you tube for more proof.

Why people are so negative everytime? I don't get it.



Mike Zarren said the corner 3-pointer was the most efficient shot BESIDES the layup. I'll use your numbers and some assumptions here:

Layup = (80%)*(2 pts) = 1.6 expected points
Corner 3 = (42%)*(3 pts) = 1.26 expected pts
Elbow jumper = (assumed 50%)*(2 pts) = 1.0 expected pts

Maybe the story could have explained this a bit more, but the editors might think most readers don't want to see the details. I'd agree with them, although I like the details.


You mentioned that the Celtics went from 24 wins to 66 in one year. Although I'm sure the number-crunching had something to do with it, I think the offseason acquisition of two superstars had a bigger impact.


To Matoli and others,

You are allowed to post whatever you want in the comments section of a blog. But, that doesn't mean you need to abuse the privilege. Your personal distaste for basketball is neither interesting nor noteworthy. Therefore, you have nothing to contribute to this discussion and should just move on and comment on stories that actually interest you. And for what it's worth, this article wasn't even really about basketball, per se, it was about personal growth.


Jose Costa (aka David Stern), thanks for writing in.

p.s. you gave yourself away by using that lame NBA tagline "Where Amazing Happens" no REAL NBA fan actually likes that slogan...

Rich Wilson

"It should be the other way around, but it's been that way all the time."

I got a great piece of advice from a very good amateur bike racer onces. He said "you will win on your strengths, so train your strengths to win, but train your weaknesses enough that they won't be a detriment".

Example: I'm not a natural sprinter. I'm never going to win a sprint against someone who is. So spending a whole bunch of energy training to sprint isn't going to win me any races. However if I'm in a small break away group of other non-sprinters, I should at least be good enough to have a chance to beat them, so don't ignore sprinting all together.

Jose Costa


I think you don't like sports at all. Amazing happens in NBA basketball. Someone who sees Kobe playing and says that's not poetry must have a prblem.

Guys, nice report. I think Ainge should aknowledge that a lot of data helps, but only top players can take a team to the championship. I.e., hiring an MIT mathmathician will not take Memphis to the finals...


Matoli, what are talking about? You just said the NBA has too much scoring and then compared football, hockey and soccer as having the same problem.. That makes no sense at all.

If you want a real franchise, look at what Seattle has done.. They have like 9 draft picks in the next 3 years, have made Durant their franchise player and will be moving to Oklahoma City (great fans if you saw the hornets play there at all for two years)


I am not sure that basketball is a subject that most readers of this blog understand. I found the article interesting, but lets not overvalue off-court decision makers. The players on the court are massively responsible for the success of a team.

One thing that I found ridiculous was that your stat maven believes that a three pointer from the corner is more efficient than a layup. A layup is is worth two points. A three pointer is worth, you guessed it, three points. Most players shoot 80% or higher on their layups. However not even the greatest three point shooters shoot over 50% from three point land. That means roughly every three pointer taken is worth 1.5 points and that every two pointer is worth 1.6 points.

Bottom line, I think throwing that comment in the article without more explanation was stupid.


Matoli, You say you don't follow the game, but then list a bunch of ways basketball is inferior. Um, how do you know that if you don't follow it?

Agree about them becoming geniuses this year. It's amazing those stats didn't stop them from having an 18-game losing skid or whatever last season.


Remember, this is the same brain trust that built that 24-win Celtics team. KG falls into their lap and suddenly they are basketball geniuses? Sorry, I am not buying it.

Now, if you want to profile someone who took a crap team and built them up in very little time through superior scouting, look at Kevin Pritchard who came to Portland by way of the dynasty San Antonio Spurs.


Nice column. The expectation might be that modern analysis overturns old ideas when most in fact validates it. When the analysis was done about stealing 3rd, which to the average fan seems a no-brainer, it turned out the stats followed exactly what the managers believed, that it's only worth doing in limited cases because of how runs are actually generated. We sometimes fail to credit the reality that human experience does accumulate to the right understanding in many, many cases. It's good to know for sure that a 3 pt'er from the corner is indeed the right choice, as basketball players have known for years.

I would be interested in seeing how the league evolved over the years while an experience base built up with the 3 point shot. The game changed over time but has been quite stable because a knowledge base, largely experiential and passed down, now exists.

You can find small edges in numbers, but for example baseball managers and insiders understood the value of walks before the average fan. Dave Kingman, for example, was known for power and strikeouts but the people in the game knew his walks, average and power not only varied a lot from season to season, which meant he was a gamble, but that he also had a relatively low number of walks. Call it OBP or construct an OPS but even without the label for the stat, the baseball people knew what they were seeing.

As for LeBron going right or left from the top of the key, remember that Joe Dumars and the Pistons Bad Boy teams came up with the "Jordan Rules" for containing Michael. They didn't have it on computer with video of each play but they played Chicago a lot and analyzed Mike's tendencies. When Ted Williams played, defenses shifted the SS behind 2nd based on what they saw and charted - and Ted, to his credit, figured out the odds of hitting into the shift versus trying to go the other way, which helped generate his absolutely incredible nearly .500 lifetime OBP. (In a nutshell, only swing at strikes, hit the ball hard and you'll walk a lot or make solid contact and a well hit ball is more likely to be a hit shift or not.)



I don't follow basketball because it is boring. Play monotonously ping-pongs back and forth, back and forth ad nauseam with very little variation. ("Celts gets the ball, they shoot, they score! Hawks take the ball, they shoot, they score!" Repeat.) It's all driven by the clocks, especially the shot clock, not any natural rhythm in the game.

The scores are often pretty close all through the game, and the whole thing comes down to the last five minutes when one team may develop a spurt of well-timed momentum or get lucky and edge out the other team at just the right time. If the NBA potentates reduced basketball games to just that last five minutes, it might be interesting.

Football, soccer, and hockey suffer from similar deficiencies, but at least the first two offer a little more variety in their game play. On the other hand, baseball, which I also don't follow any more due to the strikes of yore, allows one team to keep control and build momentum as long as they don't blow it.

No, I am not making the case for Cricket.



(Long time reader first time commenter)

Being that this is my first comment, I'm glad its on my favorite sport basketball, which I've been following for four fifths of my entire life, so I consider myself an expert. The one thing I know about basketball is that it is not the statistical endeavor that some like to make it out to be.

If you want evidence on how poorly one can create, review, or even explore the statistical field of basketball, look know further than John Hollinger of ESPN who repeatedly has Kobe Bryant low on his Player Efficiency Ratings, and this year made an argument using his PER stat that Kobe Bryant should be fourth in the MVP race. Just read the comments on any one of his articles and you would read the utter disgust from basketball purest such as myself.

Simply put, the best stats in bball are the most basic ones you can think of. Why?

1.) Bball is played by 5 guys versus 5 guys in direct competition with each other where the same 5 guys plays both offense and defense.

2.) There are an infinite number of offenses and defenses a team can play against each other, thus ruining the constant playing environment used to judge performances.

3.) Injuries, kill chemistry and can cause chaos on a team.

4.) Teams peak at different times of year. The Cavs for example always peak at playoff time when James dominates the ball.

5.) Piggybacking on that, certain superstar players (Lebron & Kobe) elevate in the playoffs and others (see Dirk) don't.



You could hire John Nash and Steven Hawkings to break it down for you, but I'll take my chances with Steve Nash and Connie Hawkins any day!

Peter Shulman

The problem with moving from baseball to basketball, I think, is that it's far too fluid of a game to isolate into meaningful statistics. With baseball, a specific batter faces a specific pitcher--nothing can change that. But in basketball, one is guarded by a team, essentially. Player A may drive by player B 80% of the time, but when player C is in to challenge the shot, he misses 60% of the time, v. his normal 60% made shots. However, player C might also give up a lot of points to his man, since he's so eager to swat someone else's shot. Many points are scored in basketball by a penetration that leads three or four passes which finally result in a basket, while the stats only show an assist for the last pass.

Now one can argue that basketball just needs to be broken down more acutely to take all of things into account. But I think it would show us precisely what close watchers already know. I'd love to see a formula, for example, that would prove what I think I know: that Charles Barkley, whose game I loved so much in so many ways, consistently failed to rotate quickly on defense for the Suns, costing them a chance at a championship. I'm almost sure that a correctly formatted analysis would show this, but that's because it's so evident (to me) that it doesn't need to be shown (and yes, I know: the Suns wouldn't have sniffed a title without all of Barkley's other gifts.)

The bottom line is that unlike baseball, basketball is terribly fluid and collective, and thus difficult to isolate individual ability outside of what can be seen by the naked eye. We can all see that Bruce Bowen, of the Spurs, messes with people's games, and by extension the entire rhythm of a team's offense. Obviously we can't all see that Jack Cust, stuck at Triple AAA for many years, could actually succeed at the Major League level, though statistics suggested that he might.