Account
NBA Insider: Is It Numbers or Talent? Sorting Fact, Fiction in NBA Stats Wave
February 26, 2015
This column is not about Charles Barkley, who recently said, "I've always believed analytics was crap."
It's not about Larry Brown, who recently opined, "To say that these analytics guys have the answer is crazy."
It's not about the purported clash of new-age analysis and old-school thinking. It's not about scouts vs. "stat guys" or some tortured debate over "gut" vs. data.
Every day, all across the NBA, reasonable people are having reasonable discussions that involve some combination of observation, judgment and statistical data. These discussions have been going on since the invention of the box score, long before "advanced stats" entered the lexicon, long before "analytics" became a movement.
The legendary Dean Smith was, as The New York Times recently noted, a pioneer in the use of advanced stats—in the 1960s. Pat Riley began compiling his own statistical formulas back in the 1980s, when he was coaching the Showtime Lakers.
Forward-thinking coaches have always explored new ways to analyze and attack the game, to test their hypotheses against any available data. There is simply more data than ever available now, and a more intensive effort to harness it.
That's what statistical analysis is really about.
But the public debate has become wildly distorted, because some loud skeptics—Barkley, Brown and countless cranky old newspaper columnists—have promoted a thoroughly warped image of the advanced stats movement, with criticism built on a foundation of straw men, misperceptions and mythology.
"It's like arguing with a baby, or someone who believes the Earth is flat," Daryl Morey, the Houston Rockets general manager and advanced-stats devotee, said in describing what it is like debating the anti-analytics crowd. "It's like debating politics on Facebook."
According to the skeptics, the analytics movement is about stat geeks who never played the game, don't understand the game and believe that math trumps everything.
Spend even five minutes talking to an analytics expert, and you will find none of these things to be true.
No one working in NBA analytics—at least no one I've ever met—claims to have all the answers. None of them believe in building a roster based on stats alone. None of them discount the value of scouting, or coaching, or chemistry. In fact, they are generally emphatic that analytics is just one tool—albeit a crucial one—in the effort to build a winning team.
And, contrary to what Barkley may believe, analytics have influenced some of the NBA's best franchises in recent years.
The Dallas Mavericks used advanced stats to make a key lineup change in the 2011 NBA Finals, when they defeated the Miami Heat. Miami, under Riley and coach Erik Spoelstra, consults the numbers regularly. Boston Celtics president Danny Ainge was an early adopter (and the one who hired Morey). The San Antonio Spurs have been quietly incorporating analytics for years.
This weekend, many of the brightest statistical minds—as well as coaches and executives—will gather in Boston for the annual MIT Sloan Sports Analytics Conference. The event is part academic exchange, part job fair, part schmooze-fest. But the attendees generally share an intellectual curiosity and a thirst to explore the games beyond the box score.
So this seemed like a good time to smash some tired tropes. Following are some of the most common myths, with responses from a cross-section of analytics experts.
Our panel:
Morey, now in his eighth season as Rockets GM.
Dean Oliver, a pioneer in the field, who now works for the Sacramento Kings after previous stops with the Denver Nuggets and Seattle SuperSonics.
John Hollinger, vice president of basketball operations for the Memphis Grizzlies and the inventor of ESPN's Player Efficiency Rating.
Ben Alamar, ESPN's director of production analytics, who previously spent five seasons with the Oklahoma City Thunder.
Aaron Barzilai, the former director of analytics for the Philadelphia 76ers and, previously, the Grizzlies.
Mark Cuban, owner of the Mavericks.
John Doe, a current team analytics director who wished to remain anonymous.
R.C. Buford, general manager of the Spurs.
Myth No. 1: Stat guys believe you can win with a spreadsheet and a calculator. Math trumps all.
Oliver: I've never said anything like that...I know what the numbers are missing. I know what the numbers don't see very well.
Alamar: There are some analysts out there on Twitter and such that actually believe that (the numbers trump everything). There's no way that's true, especially in basketball, where we're nowhere close to solving anything....It's really hard to be involved in the decision-making at an NBA team and really look at your data seriously and think that it is the be-all, end-all of decision-making. There are so many dimensions to a player and an organization, to think that the data we have trumps everything else is just ridiculous.
Barzilai: I think that most people that are thought of as "analytics guys" aren't managing by spreadsheet and would never advocate managing by spreadsheet. I think it's about information and about trying to have as much information as possible....There's always judgment. It's definitely art, as well as science.
Morey: If I were to say "former player decision-makers, all they do is talk to a guy and make decisions," that's idiotic. But that's not what they do. Even great former player GMs like Danny Ainge, they watch a lot of games, they talk to players, they look at data, they talk to coaches. Everyone is always factoring in every piece of data, whether that's the games they've watched or the coaches they've talked to, or the drill work they've put them through in a workout....All analytics is, is how do you best take that data and make the best decision for increasing your odds of winning a championship in the NBA?
Hollinger: I met one or two guys like that (who believe the numbers trump all). They don't work in the league....I don't think you can do it that way. Eventually, the same thing is going to happen to guys who think they can do it without any data at all.
Myth No. 2: Stat guys believe stats can replace scouting
Alamar: When you start talking about the "Moneyball" stereotypes, that we're replacing scouts or taking jobs away, it's just absurd. It's not possible.
Even SportVU data (the tracking system that captures every movement in an NBA game) at its best leaves so much uncertainty. We're not even close to being able to say we can replace the direct basketball performance information that scouts provide....The other stuff—personality, intel stuff that scouts do, the data will never do that.
Doe: I think that good analysis should have a collaborative relationship with scouting. Analysis can help direct scouts to focus on certain things, not emphasize certain other things as much. Even if we get the best data and analysis in the world about what happens on a pro or college or international basketball court, we're still going to need insight into a guy's character, work ethic, etc.—and scouts play an integral role in that. Analytics has changed scouting, and will continue to change it, but it's not replacing it.
Hollinger: The No. 1 straw-man argument. What, now we're just going to ignore scouting? Which nobody has ever said. It's this myth that in order to use new information, you have to discard every other piece of information that came before it. Somehow that's become the argument that gets made, when that's never been anywhere close to the reality.
Alamar: Analytics at its best is scouts and analysts working together—very complementary sets of skills. I think in most organizations, most NBA teams, scouts will still trump the data. And the way I would frame it is this way: There are decisions that are made purely out of gut instinct, no matter what the data says. I don't think there are decisions made purely on what the data says, no matter what the gut instinct says....You go into a draft room, even if the analytics group has a full seat at the table, they're outnumbered by the scouts. So if it comes down to a vote, they're going to lose.
Myth No. 3: Stat guys never played the game and don't understand basketball.
It's true that no NBA statistical analysts played in the NBA, but many of them did, in fact, play competitive basketball, including collegiately. It's also true that a majority of today's NBA head coaches never played in the NBA, either, notably Gregg Popovich and Erik Spoelstra, who faced off in the last two NBA Finals. And Ainge, who won two championships as a player, is one of the NBA's biggest advocates of analytics.
Oliver: To do it well, you want to know the game pretty well....I played in college, two years Division III (at Cal Tech). I played against Gregg Popovich's teams when he was head coach at Pomona.
Alamar: We grew up learning math and watching basketball. I'll be the first person to tell you that the scouts know way more about basketball than I ever will. But that doesn't mean that we can't have a discussion, and I can't learn from them and make my analysis better because of them....Red Auerbach never played in the NBA. I think he was pretty successful.
Myth No. 4: Stat guys think they have the holy grail, basing a player's value on a single metric, like PER or Win Shares.
Doe: That straw man is basically just a [version] of: "Is the 'stats guy' a fraud?" Someone who thinks that his metric, be it PER or RPM or another, is a useful, all-encompassing metric isn't good at analysis. Stats are great at getting insight into specific things, and you can often aggregate some things together to get useful, broader metrics. But we're a long, long way away from being able to definitively offer some sort of player value metric.
Myth No. 5: Stat guys don't even watch the games.
Hollinger: We watch a lot of (bleeping) games. I don't know where that one started. I guarantee you everyone who's doing analytics watches a lot more games than Charles Barkley does.
Doe: We care about numbers a lot more than traditionalists, but the stat guys I know watch a lot more NBA games and film than most basketball guys. Scouts and execs often watch parts of games, then chat away the rest....But I honestly don't know of a traditional basketball person who watches as much NBA action as the analytics guys I respect and/or work with. Stat guys care about using data to get insight, and what you see helps you to make sense of the data....
Yes, I watch many games. I also know that humans are quite bad at observing things and are heavily influenced by cognitive biases, and that data is an excellent tool to both overcome those things and to ascertain things that we can't readily see.
Oliver: A good scout, a good analyst, they go and they watch the game. Your eyes see the game much better than the numbers. But the numbers see all the games. And that's a big deal....They see all the detail and they really get you a lot of the story....Watching on video and going to games, those add different components to what the numbers can give you....And then there are other things that you have to talk to people about. You have to talk to the coaches.
Myth No. 6: No one is winning with analytics. Just look at the Rockets. They haven't won anything since Morey became GM.
Five of the last seven NBA titles have been won by teams that value analytics and employ them to varying degrees. As Cuban has explained, the decision to start J.J. Barea in Game 4 of the 2011 Finals, a key turning point, was largely data-driven. The Heat's decision to use LeBron James primarily at power forward, and their move to "positionless" basketball, was based on analytics. The Rockets, meanwhile, have been one of the best teams at finding hidden gems in the draft and amassing assets—the key reason they were able to trade for James Harden in 2012.
Cuban: We use it to double-check everything, from pick-and-roll to speed to you name it. It's always there as a first check against any hypothesis.
Hollinger: You don't need analytics to identify the single best player. But if you can identify a few diamonds in the rough, you're going to be way ahead of the game, especially given the salary-cap restrictions....Analytics can't replace LeBron James. Maybe I can use it to get a better chance against LeBron James.
Oliver: Do you think the Spurs have stayed on top simply because of talent, aging talent, that has been predicted to fall off the last eight years? The Kawhi Leonard pick (the Spurs acquired him in 2012) was a huge analytics pick. It was very evident in the numbers. And a lot of the tactics they use, offensively and defensively...the things they do are very analytical.
Alamar: The Spurs to me are the greatest example of this. They don't talk about it, because they don't talk about anything. Look at that team that won the championship (last year): Who exactly is the guy who is an MVP candidate? Tim Duncan is a great, great player. But going into last year, did anyone think that was a team so loaded with talent they were going to win the championship? No.
Myth No. 7: The Rockets have built around superstars James Harden and Dwight Howard, proving it's about talent, not stats.
Or, as Barkley put it, "The Rockets sucked for a long time, so they went out and paid James Harden a lot of money. They got better. Then they went out and got Dwight Howard. They got better." Key point: Few people in 2012 projected Harden, then a Thunder sixth man, as a future MVP candidate.
Alamar: Well, when [Morey] got Harden, Barkley himself thought that Morey was overpaying for him. Analytics help you identify talent and to project forward who is going to be that superstar. When we drafted Russell Westbrook in Oklahoma City, a lot of people said we were dumb....Now he's being talked about as an MVP candidate. The analytics helped that (choice).
Morey: Charles is right, it is all about getting high-end talents. It's about recognizing them, which not any idiot can recognize....The Harden trade was about 60-40 in favor of our side at the time. Even we had doubts. You need to prepare yourself to be ready for those opportunities when they come. It's not like anyone's giving away these good players. You have to maneuver your way into them....The notion that people in analytics don't think talent wins is ludicrous.
Myth No. 8: Coaches hate analytics.
Barzilai: Analysts are working with front offices, but you could probably argue that coaches are kind of applying it more on a day-to-day basis....More and more coaching staffs are looking at what people think of as advanced metrics now—lineups, passing stats.
Hollinger: We've seen it a lot with lineup data, with things like play calls, play-call efficiency....Trades and signings were the first frontier, then the draft. Now I think it's moved a little bit beyond personnel and into the day-to-day....Even going back to Riley and Jim O'Brien and those guys, when they were coaching 15, 20 years ago, they were using a lot of data in what they did. It wasn't as advanced. Those tools didn't exist then. Hubie Brown, when he was here in Memphis, he had his game preparation sheet, with all these statistics. It's still a template a lot of teams use.
Myth No. 9: Analytics is a strategy for winning.
Alamar: Analytics is not a strategy. Analytics is a set of tools that can be used to support a whole range of strategies.
Buford: The people that attack the use of analytics are ignoring the fact that, once you keep score, once you pay people for the amount of points they produce, the amount of rebounds they produce, once you recognize that that's of value, as a comparison tool, then you're recognizing that they are using analytics. They may not want to espouse advanced metrics, but they are keeping score. I don't think analytics is a magic formula; it's an approach.
Around The League
• George Karl's first power move as Sacramento Kings coach was to trade for Andre Miller, a trusted old hand from his Denver days. Expect more of the same this summer. Sources say Karl would love to acquire Ty Lawson—if the rebuilding Nuggets were to make him available—or any other members of his last Nuggets team, which won 57 games in 2012-13. Karl wants players who move the ball and push the tempo, and that could mean wild upheaval on the Kings roster. Even DeMarcus Cousins, their franchise center, is not untouchable, according to a source with insight into Karl's thinking. "At the trade deadline, everyone was available," the person said. Including Cousins? "Every single person (on the Kings roster) was available." Although Karl does not hold a front office title, "he definitely has control" of future personnel decisions, the source said.
• The NBA trade deadline turned into homecoming day, with Kevin Garnett returning to Minnesota and Tayshaun Prince going back to Detroit. Could an Amar'e Stoudemire-Phoenix reunion be next? Perhaps. Stoudemire joined the Dallas Mavericks last week after being waived by the Knicks, but he'll be a free agent again in July, and sources say he would welcome a return to Phoenix, where he spent his first eight seasons. Stoudemire loves the city, and his creaky knees could benefit from a reunion with the Suns' renowned medical staff. Whether the Suns would reciprocate the interest is unclear. But the Suns could use a scoring big man, and team officials could certainly use a PR boost after trading the popular Goran Dragic to Miami last week.
One-Question Interview
Shawn Marion, a four-time All-Star and NBA champion (with Dallas in 2011), announced in January that he would retire at season's end, after 16 seasons. The decision was as much emotional as physical: Marion, 36, became a father last April and, as he tells Bleacher Report, his priorities have changed. Here, he explains how he knew it was time to retire. (Quotes are condensed for brevity.)
"Most guys ain't prepared for it. It depends on where you're at in your career. What kind of career have you had? Are you content? It's about where you're at in your life, actually. Where you're at in your personal life, as well as your finances. A lot of guys keep playing because of finances, mostly." (Note: Marion has earned about $150 million in salary.)
"The camaraderie, though. That's the biggest part most guys miss. Like being around the guys, just (messing) around and talking (trash). I've developed relationships with a lot of guys, and it will continue; it won't be the same as being in the locker room, being on the road and stuff."
"I have a newborn son now (Shawn David Marion), and it's a little hard when you're not together, not with your son's mother. And you're away from him for periods of time, and you're missing everything. You're seeing everything from videos and stuff. Especially at my age, it's hard....It's like, OK, you got choices to make, decisions to make, about your family and life."
"He just started walking. And I missed it. I was able to see him for a couple days, and the next day he started walking. So I had to see it from video. It was great—I was geeked up, I was with my friends, we was running around, going crazy, doing somersaults. I'm like, 'My boy is walking!' It's crazy."
"The older he is, the more I want to do stuff (with him). I see guys, I see my (friends), they talk about their kids, and doing this and going to camps. That's stuff I'm excited for."
"I told myself when I first came in, I wanted to play 15 years. And that was it. So I exceeded myself. Your body starts breaking down, and minor injuries become lingering. It's that much harder. And I played at such a high level for so long, I feel like if I can't continue to do that, my body's telling me just to shut it down. I'm not saying I can't play ball, but not play at this level.
"I've been fortunate enough to play this long and be able to do it very efficiently, and I'm truly blessed. I don't take [any] of this for granted. I've been able to provide for my family and the future. It's just about sustaining this. It's been pretty awesome."
"I just want to go out with a bang. Go out with a bang, man. And have fun. These days, you can't put a price on this stuff sometimes. It's like, man, you know it has to come to an end sooner or later. You can't play ball forever....It's time."
Insider's Take
A scout evaluates new Pistons point guard Reggie Jackson, who was acquired from Oklahoma City at the trade deadline last week. Jackson was immediately installed as the starting point guard, replacing the injured Brandon Jennings. Jackson is a restricted free agent this summer.
"Earlier in the year, when [Kevin Durant and Russell Westbrook] were hurt, he enjoyed being 'the guy.' Even though their team was bad and not really having much chance of winning games, he seemed to enjoy that. As players came back and his role changed, you could see his zest for playing had shrunk....He felt like he was the next James Harden, and he needed to be freed so he could prove that to everyone. And he got his wish.
"[The Pistons] seemed like they had really good chemistry even after [Jennings] got hurt, with D.J. Augustin playing the point, and they were making a push toward the playoffs. It will be interesting to see how locker room dynamics change there. Greg Monroe is in a similar situation, and he wants to make sure he gets his points and rebounds going into free agency, so he can shoot for that elusive Roy Hibbert money.
"Reggie's a talented scorer. He is a capable passer—but that's a secondary thought. He's more likely to try to get the assist pass than the secondary assist, or hockey assist. He seems to have an arrogance about the way he plays. He's a good defender, good on the ball, gets good pressure, steals when he's off the ball, gets in the passing lanes. He's definitely a talented player.
"They're putting [Jackson] in position where he's got to be a star. But you've got Andre Drummond, who needs the ball, and somebody [needs] to get him the ball. If you keep Greg Monroe, then he needs somebody to get him the ball. [Coach Stan Van Gundy] proved to be a magician in that before Brandon Jennings got hurt. [Jennings] was a loved teammate and playing some of his least selfish basketball of his career. Playing [Jennings and Jackson] together is feasible from a basketball standpoint, but from the two personalities standpoint, and the fact they both have to have ball in their hands, I think it's another issue."
Howard Beck covers the NBA for Bleacher Report. Follow him on Twitter, @HowardBeck.