Nerds, numbers and the new NBA

In this special edition of Coast to Coast, J.A. Adande discusses the growing analytics movement in the NBA with Tom Haberstroh, who often writes about basketball analytics for ESPN Insider. Haberstroh is sitting in for Israel Gutierrez for this week's Coast to Coast.

J.A. Adande: By most accounts, the war over NBA numbers is over, and the analytics guys have won. My thing is ... what exactly have they won? Just because advanced metrics can be found everywhere, and even creep into conversations with players, that doesn't mean they've achieved the ultimate conquest. Last I checked, victories in the NBA are measured by Larry O'Brien trophies. And those still tend to go to the teams that amass the most outstanding players, the ones who don't need numbers to demonstrate their worth.

The most essential computations that formed the two-time champion Miami Heat were the number crunches that went into figuring how to fit LeBron James, Dwyane Wade and Chris Bosh under the salary cap. And I'd argue the biggest reason the Dallas Mavericks kept them from three-peating wasn't analytics but old-fashioned faith: Mark Cuban's unwavering belief that Dirk Nowitzki could be the best player on a championship team even after all of Dirk's previous playoff shortcomings.

Even then, the Mavericks wouldn't have won if Oklahoma City Thunder general manager Sam Presti held on to Tyson Chandler. Presti was leery of Chandler's injury history and backed out of a proposed trade with New Orleans, which led to Chandler being in play for the Mavericks in 2010. Presti is considered one of the more forward-thinking GMs, but you might have noticed the Thunder have remained one of the league's best teams while insisting on playing Kendrick Perkins, a guy who makes statheads scream because his lineup efficiency numbers are so poor.

Meanwhile, Sir Stathead himself, Daryl Morey, is enjoying his greatest success so far after going the most traditional route in NBA history: signing an All-Star center. Yes, the Houston Rockets still employ the trendy layups-3s-free throws approach. It just seems to be working better now that they have an old-school big man in the middle.

Tom HaberstrohInsider: First, I don't really ascribe to the notion that analytics have won anything. The scouts versus stats battle is an imaginary one fit for a Hollywood screen, not rooted in reality (though I will say I'm cool with Jonah Hill repping the geeks in "Moneyball." Jonah Hill is awesome).

I guess if we're going to point at one particular victory -- and I don't know if you can even call it that -- it's that analytics have earned a seat at the table. We wanted to be part of the conversation, because we feel that information or empirical data is useful to making decisions. Information is power.

I get your point about Larry O'Brien trophies, but the dirty secret is that sports are decided by statistics. I know! Those pesky statistics trying to ruin all the fun! But really, points scored and points allowed determine the outcome of games, which determine the outcome of playoff series, which determine the eventual champion.


There's a reason why Baskin Robbins flourished with 31 flavors. There are options. ... Not cool with pace-adjusted stats? No worries! Plenty of other delicious flavors are out there.

"-- Tom Haberstroh

Statistics are at the heart of it all. The Dallas Mavericks may have been helped by their faith in Nowitzki, but how do you know that their conviction wasn't analytically-driven? Before their title, the peanut gallery killed Dirk for being "soft" and "not a winner," throwing him under the bus for weird, contrived intangibles. You know who didn't believe in that stuff? The numbers gurus. Mark Cuban employs one of the largest armies of data in the sport and actually had "stats guy" Roland Beech, founder of 82games.com, on the bench during the Finals helping Rick Carlisle exploit the Heat's wonky lineup choices. (Seriously, was Erik Spoelstra still starting Mike Bibby?)

Maybe it was analytics that helped Cuban build around Nowitzki and find the right complement to his unique skill-set. It doesn't take a Ph. D. in mathematical modeling to know that Dirk is awesome. But which players can get the most out of Dirk? And are they properly valued on the market? These are questions that analytics can help solve. And let's get something clear: Analytics shouldn't be the only voice in the room. But it's a voice that should be heard.

To your point about the Rockets: Maybe they don't get Howard without slowly accumulating undervalued assets that can be used to get a young star like James Harden, who can then be your biggest recruiter of superstars like Dwight. And analytics can help you identify and exploit those building blocks -- draft picks, efficient players like Kevin Martin, etc. -- before they become properly valued.

As for Perkins, I'd posit that the Thunder are winning despite Perkins, not because of him. If you're a GM, would you rather not know that the team does far worse with Perkins on the floor? See, I'd want to know that information.

Adande: Information is a good thing, and it's fun to have the additional information at our disposal these days, such as "hockey assists" (or secondary assists, as they're formally known), assist opportunities and points created by assists. One of the reasons I like the advanced assist stats so much is they tell us more about player interactions, which was a large limitation of the traditional numbers. We're getting better defensive measurements as well.

The problem is there can be a point of too much information. I've heard several coaches say they're still trying to figure out what to do with all of these numbers, still determining which ones merit passing along to their players. Then there's the matter of how much attention will the players pay to these stats? We know a Shane Battier or Arron Afflalo might delve into the data, but do you think anyone volunteered for the job of telling Perkins that the numbers suggest he should take a seat?


Are we wasting time trying to find the killer stat when it already exists? It's like these hotels that try these cute phrases to put on their door tags, when nothing gets the point across better than 'Do Not Disturb.'

"-- J.A. Adande

Roy Hibbert has reluctantly gone along with what the Indiana Pacers' stats staff told him: The best thing a post player can do with the ball is pass it out to a teammate for an open 3-pointer. Not every player would be willing to do the dirty defensive work that Hibbert does without the reward of getting post-up plays run for him. That's one of the challenges of taking stats from the reflective -- which they're better at than ever -- to the predictive. There's a risk in assuming that mathematical models will work in the real world, where egos, interactions and injuries all have to be taken into consideration.

Sometimes the pursuit of perfect data leads us down the wrong path. I understand the need to extract individual and group performances to measure a standard time frame. But do we really need to know team offensive and defensive stats over, say 100 possessions? Aren't points scored and points allowed per 48 minutes the only stats that really matter?

I get your earlier point, that we need stats to keep track of who won and lost. But sometimes the efficiency of how they do so is beside the point. The irony is the pursuit of efficiency data can actually become inefficient; sometimes it's clutter.

I read a blog post once that broke down the efficiency numbers of a team's loss, putting the percentages into context. The final score was something like 110-80. That alone indicated their defense and offense were both bad.

Scoring margin tells me all I need to know. It can account for stat-warping pace differences by focusing on the essentials: Are you outscoring your opponents on a regular basis? The top five scoring margin teams this season are the Pacers, Thunder, Clippers, Spurs and Heat. I'm pretty confident the NBA champion will come from that group.

In fact, nine of the past 10 champions have ranked in the top five in scoring margin during the regular season (the exception: the 2011 Mavericks). Are we wasting time trying to find the killer stat when it already exists? It's like these hotels that try these cute phrases to put on their door tags, when nothing gets the point across better than "Do Not Disturb."

Haberstroh: Totally hear you about the T.M.I. thing. Just the other day, former Rocket Toney Douglas sat at his Heat locker pregame, flipping through the pregame scouting report packet that the Heat provided. He turned to fellow former Rocket Battier and said, "Man, this is so much better than Houston's book of numbers." And Shane laughed and replied, "For me? Nah, give me all the data. Can't get enough."

The moral of the story: One size does not fit all. And that's true in any walk of life. You might have had a professor who really struck a chord with you in college who made others doze off. You might have read a book that you couldn't put down, but the next guy at the library read two words before he grabbed another one off the shelf. Not everyone learns or consumes information the same way. Not everyone can be Shane Battier.

But what's interesting to me is how many people tell me, "Man, I never thought to look at the game this way. I didn't even know these numbers exist!"

Numbers are not for everyone, but for a large subsection of folks, they've never peeked behind that digital curtain. For many, it's about discovery. There's a reason why Baskin Robbins flourished with 31 flavors. There's not just chocolate and vanilla anymore. There are options. And no one's forcing anyone to choose. Not cool with pace-adjusted stats? No worries! Plenty of other delicious flavors are out there.

You're on to something with the communication concerns. Teams are still wrestling with this just like they are grappling with players on social media. Technology is helping us move forward faster than ever and there will be bumps along the way. But when it comes to Hibbert, getting players to do what's best for the team and not necessarily what's best for the individual isn't an issue unique to analytics. Teams want to win and, at the end of the day, it's up to the coaching staff and management to digest and leverage the information the best way possible to win. Maybe they have to go about conveying that information a different way. Yes, egos are going to be adjusted. But hasn't ego management always been a part of the coach's job description? If you're hurting the team, you're hurting the team.

Adande: While we're on the subject of egos, never underestimate the ego of coaches and general managers who think their programs can fix players whom the numbers indicate can't get it done. Sometimes stats can provide irrefutable evidence.

And, yes, sometimes the numbers provide illumination, explaining why a team is succeeding or failing, or how a player can best be used.

I stop short of saying numbers provide enjoyment. I watch the NBA for the athletic feats, the dramatic moments and the quirky personalities. They have yet to find a way to quantify those. So the advanced stats still haven't won me over.

Haberstroh: I think there are two things going on here. Basketball is a game decided by numbers, played by humans. We all watch basketball because it's a beautiful game that showcases humans doing awesome superhuman things. But there are fans who also watch it because they buy the jerseys and invest so much capital (in both time and money) in wanting to see their team win, wanting to be a part of success. And understanding why wins or losses occur, or how a team might improve itself, can make the fan experience that much more satisfying.

They want to know: Why did that happen? Was that the right thing to do there? Analytics can help answer those questions because the sport is inherently statistical. In the end, basketball is not art, nor science. It is both.