Money Ball in the NBA
November 24, 2013
By Bruce Marshall
Making Sense of Hoop “Analytics”
Watching a practice session of the 1984 Olympics team he was coaching, Bob Knight was sitting in the stands with Stu Inman, an old friend and then the general manager of the Portland Trail Blazers. The NBA Draft was to be held in a few days, and Inman's Blazers were picking second. It was general knowledge that the Houston Rockets, picking first, would take University of Houston center Akeem (before he added the "H") Olajuwon.
Following, courtesy of the excellent biography entitled Knight, by Bob Hammel, is Knight's recounting of the brief discussion that followed with Inman, who had just told Knight that the Blazers were leaning toward taking Sam Bowie of Kentucky.
"Stu," I said, "you've got to take (Michael) Jordan."
"Bob, we need a center."
"Well, play him at center! Nobody could guard him. He's the best player there is. You have to take him!"
Of course, we know what happened a few days later. Portland took Bowie, and the Chicago Bulls, drafting third, took Jordan.
We can assume, 29 years later, that the Bowie-over-Jordan pick in the '84 draft had something to do with the basketball analytics of the day. Those calculations have changed a lot in the three decades since. But then, as now, and as Bob Knight would likely remind us, great players make great stats...not the other way around.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Business school meets the hardwood. Hoop-style "Moneyball" in action. Call it whatever you wish, but a statistical revolution is in the process of turning basketball, at both the college and especially the pro level, inside out.
Whether referred to as "basketball analytics" or "basketball metrics" or whatever, it is at its base an absolute reliance on statistical data and formulations that has altered the look of many NBA front offices and spilled into roster construction and coaching hires. For many pro basketball organizations, the new phrase might as well be "get on board or get out of the way."
Before going any further, it should be noted that most of this "new" statistical data is not really that new at all. For many years, NBA teams have had access to this sort of "advanced" information, including the "plus/minus" player and lineup calculations that are at the root of much of the modern stat revolution. And, in many cases, basketball coaches have been information junkies for years, dating from legends such as Hank Iba and Dean Smith.
Although the new "metrics" have evolved to the point where teams rely on statistics that didn't exist a few years ago. Such as effective field goal percentage (which weights three-point attempts differently), points per possession (rather than points per game), rebounding percentage (the percentage of rebounds a player grabs while on the court), and Player Efficiency Rating (or PER, a catch-all stat that aims to capture a player's overall production; more on this is a moment).
The "new wave" hoop thinkers, however, have taken the stat calculations to new levels and have now infiltrated the top management ranks of teams more than ever before. In a sense, it mirrors the "Moneyball" dynamics in baseball most identified with Oakland A's GM Billy Beane, who was the first MLB exec to rely heavily on an interpretation based upon an advanced statistical model to build a roster and organization to formulate the "A's way" to play.
More than a decade after Beane shot to prominence in Oakland, several Beane-like clones have surfaced in MLB. But it was no overnight phenomenon in baseball; it took MLB a while to adapt to Beane-style thinking (and several holdout organizations still remain). But the stats revolution seems to have made a more dramatic breakthrough in the NBA, where many organizations have adopted the "metrics" model in quick order.
One of the backers of new-style "hoop metrics" is incoming NBA commissioner Adam Silver, whose interest in analytics dates from his days as a law student at the University of Chicago. "Clearly now, throughout the league, there is a cross section in terms of how they value it," Silver said in a recent interview with the Washington Post, "but I think there is a sense now from every team that at least it's a factor, in considering lineups and considering players, and some teams use it more than others."
More than two-thirds of the NBA's teams now have full-time staffers charged with dissecting numbers. In many of those cases, the resulting data have changed the way general managers construct rosters, coaches teach players, and how those players act and react on the court.
For those teams, every off-court decision is considered in terms of risk, reward and value: trades, draft picks, shot selection. For example, number-crunchers all agree teams should shoot more threes--high risk, but also high reward. The result? NBA teams averaged 20 three-point attempts per game last season, up 26% from a decade ago and more than double the mark from 20 years ago.
No NBA team symbolizes this "new wave" of hoops thinking any more than the Memphis Grizzlies. Considering the changes in the front office and bench in the past year, and in the offseason, one might think the Grizzlies were undergoing a transformation borne from necessity, rather than a franchise-best regular season and playoff run in the 2012-13 season under HC Lionel Hollins.
The changes in Memphis began with an ownership switch that transferred the Grizzlies from Chicago-based Michael Heisley to a large group of investors led by young (35 years old) Silicon Valley tech entrepreneur Robert Pera, a former Apple hardware engineer who eventually went out on his own in the wireless arena and founded Ubiquiti Networks, an endeavor that would make him a billionaire when the company went public in 2011. Subsequently, Pera was at the vortex of an investment consortium that would eventually purchase controlling interest in the Grizzlies.
This is the same Pera who would challenge Michael Jordan to a one-on-one basketball matchup shortly after purchasing the Grizzlies. To his credit, Jordan merely smiled and politely declined the offer. Although we don't know Pera, we do know of many young tech hotshots who, in the same situation, would likely believe they had won a sort of stare-down with Jordan, inflating their egos further. Judging from what we hear, it would not surprise us if Pera could be counted among those sorts.
The Pera group would appoint a new team CEO for the Grizzlies, Jason Levein, a former player agent and assistant GM of the Sacramento Kings in what would be another part of the management overhaul for the Grizz. Also replaced was most of the scouting staff with another new group led by new GM John Hollinger, a former ESPN.com writer and a "pioneer" in the advanced stats movement.
(Hollinger is credited with developing the aforementioned Player Efficienty Rating, or PER. While these numbers validate that some superstars, such as LeBron James or Kevin Durant, are good at what they do [what a surprise!], they also reveal that others, such as Rudy Gay, Brandon Jennings and Ricky Rubio, might not be as impactful as the traditional box score suggests. Such metrics, coupled with salary cap issues, were factors in last winter's trade of Rudy Gay, Memphis' leading scorer and most popular player, to Toronto. Eyebrows were raised around the league, and Hollinger declined to specifically address Gay, but said in general, "I think there's a better understanding of what is a high-value player or a high-value shot versus just what looks good." )
Hollins and Hollinger, however, never quite saw things the same way, which surprised no one who knew either. Especially Hollins, who typified an "old school" approach to hoops hardened through his years as, first, a star player at Arizona State, then a decent NBA playing career, followed by a long and meandering path through the coaching ranks that would eventually land him the head coaching job in Memphis.
Hollins viewed hoops in simpler terms, preferring to put his best five players on the floor, then coach to their strengths. With the Grizzlies, that meant relying upon bigs Marc Gasol and Zach Randolph and an inside-out offense that would pound the ball into the paint in halfcourt sets, stress sharing the ball, control tempo, and defend with zeal. Using this strategy, most NBA observers believe Hollins extracted almost all he could from his team, which won 56 games in the 2012-13 regular season and made a first-ever franchise trip to the Western Conference finals.
Not good enough, however, for new GM Hollinger, who did not believe Hollins was on board with the new stats-based philosophy. He was right; Hollins wasn't on board. Hollins was thus made the scapegoat for last spring's playoff sweep administered by the Spurs, and was summarily dismissed. In his place would be assistant Dave Joerger, promoted to Hollins' old job and an enthusiastic proponent of the Hollinger "metrics" approach.
About the same time, and somewhat similarly, longtime HC George Karl was also dismissed in Denver, effectively another victim of the new "analytics" revolution.
A lot of NBA observers looked at the developments in Memphis and realized that a fundamental change was beginning to take place in the league and front offices. In many cases, these were not going to be "basketball" people in charge of the personnel operations. "Techies" (nerds?) were going to be calling the shots in a variety of locales.
Another trailblazer (and we don't mean the Portland variety) in hoops metrics is on display in Philadelphia with the Sixers and their new, 35-year-old GM Sam Hinkie, a former Houston Rockets assistant GM and a Stanford School of Business graduate who cut his teeth at Bain & Company, a management consulting firm. Hinkie is now presiding over what promises to be either one of the league's worst teams or one of its greatest experiments.
Or, perhaps, both.
Sources tell us that Hinkie is likely to applaud a Sixers' missed three-point shot from the corner that ends up as an opponent rebound, while groaning when a Philly player connects on a semi-contested mid-range jumper. Hinkie's reactions are because he watches the game and judges every action based on probabilities...what should have happened, not necessarily what does happen.
Such is the credo of the new NBA, and what happens with the Sixers in the next couple of seasons could be the truest test yet of the analytics movement. Hinkie hopes he can build a team from the ground up, aided by formulas, statistical models, and data-centric philosophies that will reduce risk and lead to smarter decisions.
(Early results for the Memphis and Philadelphia experiments are mixed this season, with the Grizzlies perhaps the league's biggest disappointment in the early going, with some sources suggesting the players detest the new direction of the organization and are simply not playing as hard for Joerger as they did for Hollins, while the Sixers have been a pleasant surprise and hardly as bad as many envisioned.)
Another angle of the new "basketball metrics" was presented at the Sloan MIT Sports Analytics Conference this past spring.
The concept? The five normal positions of point guard, off guard, small forward, power forward, and center should be replaced by these ten new positions that each successful team must have.
Ten New Positions
Two-Way All Stars...Think Kevin Durant, Kobe Bryant
Inside/outside scorers...Think Jeremy Lin, Chandler Parsons
Jump-shooting ball handlers...Think Damian Lillard, Steph Curry
Defensive ball handlers...Think Ty Lawson, Kyle Lowry
3-Point ball handlers...Think Steve Nash, Ray Felton
Low-usage ball handlers...Think Trevor Ariza, David Lee
3-Point specialists...Think Steve Novak, Shane Battier
Mid-Range big men...Think Brandon Bass, Big Baby Davis
Scoring rebounders...Think Tim Duncan, DeMarcus Cousins
Paint protectors...Think Larry Sanders, Joakim Noah
Metrics-centric types such as Hollinger and Hinkie and many others are developing their teams based upon these new positions. Although many savvy old-school hoops observers are not impressed, suggesting all these new positions redefine are the roots of basketball, still leaving two guards, two forwards and a center. There is no fundamental change, many oldtimers suggest, from the 1950s and 1960s.
But application of these new-style hoop metrics has indeed altered the way the game is played in the NBA. Scoring has moved closer to the basket (low risk) and beyond the three-point arc (high reward). Players whose bread-and-butter has been a midrange jump shot increasingly find themselves on the bench...or out of the league. For example, a star player from the '70s like former Bull Bob Love, who lived on the mid-range jumper, might find it more difficult to locate a situation where he could flourish in such a manner today.
Long-time NBA observers have been quick to note this fundamental change in the game, which has trickled into the college ranks as well. In the pro game, while three-point shots are up dramatically, the number of long two-point attempts (15 to 21-footers) keeps shrinking, accounting for nearly 22% of all shots as recently as five years ago but reduced to just 18.5% of all shots last season. Meanwhile, shots at the rim (three feet or less) have steadily increased, now accounting for one in three shots. Such numbers roughly translate to the college game as well.
We've watched enough hoops, and have had contact with enough coaches, scouts and observers within the game, to view this rush to change with some suspicion. Indeed, we believe one of the best approaches to the new hoops analytics is provided by Miami Heat HC Erik Spoelstra, who has had plenty of exposure to hoops metrics throughout his career and essentially believes that analytics are part of the equation...not the equation itself.
Some of our hoop contacts are not convinced about the new-wave thinking in hoops, either. What we suspect, as do some of our most-trusted insiders, is that the hoops metrics revolution appeals to a lot of the new-wave owners who want to apply business-school theories to hardwood. This ilk, like Robert Pera in Memphis, do not relate to the many former players who have long held personnel management positions in the league; rather, they talk the same language as sorts such as Sam Hinkie and John Hollinger, who have seized this opportunity to begin moving out the old-school types and ex-players who have often swelled the league's coaching and administrative ranks. Many new owners will relate more to Sam Hinkie types than Jerry West.
We think it is worth noting that another devotee of analytics, Houston GM Daryl Morey (the mentor of Hinkie), was close to losing his job a year ago before Oklahoma City threw Morey and the Rockets a lifeline by trading James Harden. Mostly because of Harden, Morey's job was saved last season...and thus one of the great proponents of analytics not only rehabilitated his reputation, but also was able to add extra buoyancy to the hoops analytics movement.
By us, basketball remains a game of playing together, playing smart, playing with skill, and playing hard. Good luck in measuring every aspect on a spread sheet. We have even viewed the standard bearer of analytical stats, the "+/-" calculation, with some suspicion for many years. One bad rotation can distort such numbers, which even at their most illuminating was always going to be static, as are all of the basketball metric calculations (and all stats, for that matter) that are simply finding various ways to measure past results. The circumstances that created those numbers, however, are always bound to be in flux.
We're not completely anti-metrics, however; indeed, we believe the one stat calculation that is particularly illuminating is the 3-point shots made + free throws made number. And the old "+/-" and many of the Hollinger and Hinkie-favored equations have some merit. Although, in the end, the one calculation that works 100% of the time is simply scoring more points than the opposition.
Which, not to sound flippant, serves as a reminder that stats and hoop metrics are often back-fitted and after the fact; some of our most-trusted hoops insiders suggest that sorts such as Hollinger view their calculations as the basis for stats, rather than vice versa.
Moreover, defensive calculations seem to have been mostly overlooked by the Hollingers and Hinkies of the NBA. When recently challenged about including hardly any defensive numbers in his calculations, Hollinger begrudgingly added in categories of blocked shots and steals to another PER calculation, while grousing that neither stat was meaningful.
(Which, by us, suggests that types such as Hollinger have no practical experience with the game; ask anyone who has played how unnerving it can be to have a feared shot-blocker like Bill Russell roaming in the paint or elsewhere on the floor, or how a Muggsy Bogues-sort can distort a game on the defensive end with an innate ability to steal the ball.)
Pull up a chair, this storyline is just starting to get interesting.
10-3 L13 G-Plays, 12-5 L17 Picks
4-0 L4 Guarantees, 6-2 L8 Picks
12-4 L16 Guaranteed Plays
9-2 L11 Picks, 7-2 L9 Guarantees
15-5 L20 Picks, 12-4 L16 G-Plays
21-11 L32 Over/Under Plays
With a VI Gold Membership, you can SAVE 20%
off Daily Pick packages, SAVE 10%
off a Live Odds Pro subscription, and receive access to up to 1,000
Member Plays each month!