CHAPTER 3


What Gets Measured Gets Noticed

For Justyn Ross, the numbers started to add up. As the top recruit in the 2018 high school class of football-crazed Alabama, he was attracting not only the state’s two dominant football programs, the University of Alabama and Auburn University, but over twenty more Division 1 powerhouses, including 2017 NCAA champion Clemson University. As a five-star recruit and a 2018 Under Armour All-American, he ranked as one of the top five wide receivers in his class and in the top twenty of all positions in the country. He was certainly one to watch when he arrived at Buford High School, about an hour northeast of Atlanta, on a warm March Sunday for one of the thirteen regional events of Nike Football’s “The Opening,” an annual showcase of the top prep players.

With almost 500 local playmakers at each event, getting noticed enough to be one of only 166 players invited to the national final in June at Nike headquarters in Oregon requires both a football résumé of achievement along with stellar performances at the regional event. In addition to position-specific drills and competitions, all recruits have their raw athleticism assessed through the four tests that make up the Nike Football Rating. Scores on a forty-yard dash, twenty-yard shuttle run, a vertical jump test, and a kneeling power ball toss combine to create the single rating metric, also known across sports as SPARQ (Speed, Power, Agility, Reaction, Quickness).

As a six-foot-five-inch, 195-pound junior starring for Central High School in Phenix City, AL, Ross dominated his opposition, catching 38 passes for 663 yards and 8 touchdowns. “He’s just a very coachable kid and probably one of the best athletes I’ve ever been around, period, in my 23, 24 years of coaching,” said his head football coach Jamey DuBose.1

To be sure, Ross’s impressive physical stature at age seventeen is usually enough to put fear in local opposing defensive backs, even in Alabama. But what about when going up against the top defenders from across the South? To be a playmaker on that stage, there has to be more than just speed, power, and agility, since everyone comes to the table with those tools. Like the other players at the Atlanta regional, Ross put up impressive numbers in the four SPARQ tests: a quick 4.87 seconds in the forty-yard dash, 4.40 seconds in the shuttle run, 34 feet in the power ball toss, and 27.3 inches in the vertical jump, for an overall rating of 74.2.

However, of the forty-nine wide receivers at the Atlanta event, Ross’s SPARQ rating came in at forty-first place. The five-star recruit could not even place in the top half of his position group in pure athleticism. In the two key measures for a wide receiver, the forty-yard dash and vertical jump, he ranked thirty-seventh and forty-sixth, respectively. Just a bad day of testing? Possibly, until he finds out at the end of the day that he is one of only six players total, out of the entire field of more than four hundred, to be invited to the national event in Beaverton, Oregon.

“It means a lot to me to be invited,” said Ross, who has received more than twenty Division I scholarship offers, “because I’ll have a great chance to showcase myself against the best of the best across the country.”2

Ultimately, Ross signed to play for Clemson University, marking the first time the top-rated high school player in Alabama went out of state since Jameis Winston left for Florida State in 2012.

There were other mismatches at the Atlanta event: under-recruited players with high test scores and vice versa. Channing Tindall, a three-star-rated linebacker ranked thirty-seventh-best nationally at his position, was invited to Oregon, thanks to his regional-winning SPARQ score of 125. However, of the top ten SPARQ scores at the event, only Tindall will be headed to the finals.

On the other hand, another finals invitee, Jamaree Salyer, a five-star recruit and the sixth overall ranked 2018 player in the country, only placed fourth in SPARQ among the offensive linemen and seventy-first overall at the Atlanta event. Ja’Marcus “J.J.” Peterson, a four-star linebacker also invited to the finals, placed fifth among his position group and twenty-sixth overall. Tindall and Salyer will be teammates at the University of Georgia while Peterson signed with the University of Tennessee.

There’s no question that all of these players are elite athletes, but to be a playmaker who attracts the interest of top colleges requires something more. Otherwise the highly coveted recruits would consistently place among the fastest, strongest, and quickest. Physical strength and conditioning is a must for staying on the field, but being the champion of the weight room does not guarantee a spot in the starting lineup. Take away pure athleticism and you’re left with cognitive-based, sport-specific skills to differentiate the elite players from the better-than-average crowd.

Still, for some coaches, the allure of the physical specimen with the raw tools is enough to take a chance on, especially at the pro level. Because when all else fails, what can get measured gets ranked, and what gets ranked can justify a draft pick. And that’s the allure of the SPARQ rating.

As players progress from high school to college to pro, the talent pipeline filters out all but the best athletes. The marginal differences in physical attributes as well as in-game performance are measured in decimal places rather than whole numbers. Coaching staffs feel confident that they can teach the finer points of the game to these uber-athletes. In football, the drafting philosophy known as “best player available” is a nod to the well-publicized NFL Scouting Combine physical tests, a modified set of SPARQ standards.

According to Zach Whitman, creator of the 3 Sigma Athlete3 blog, the dream is to find the freak athlete buried somewhere in an underappreciated and lesser-known college conference who can stand up physically to the rigors of the NFL. Whitman has analyzed hundreds of players, comparing a prospect’s position-adjusted combine results, what he calls pSPARQ,4 with their counterparts already active in the NFL. This comparison, known as the z-score, will be 0.0 if the prospect matches the league average for his position. A z-score of 1.0 indicates a prospect who is one standard deviation higher than his position peers. A negative z-score warns coaches and GMs that the player will enter the league as a sub-par athlete, relatively speaking. Whitman’s blog name, 3 Sigma Athlete, honors the five prospects who ranked at least three standard deviations (three sigmas) above their NFL peers when they were drafted: Evan Mathis, Byron Jones, Calvin Johnson, Lane Johnson, and J. J. Watt. Being in this exclusive club places you in the 99.87th percentile of NFL players.

As Whitman likes to say, “Not all good athletes are good players. Very few poor athletes are good players. Most great players are great athletes.”5 In a 2015 study,6 to demonstrate this at a macro level, he plotted the pSPARQ score of every NFL player from 1999 to 2012 against another new age metric, approximate value (AV).7 Created by Doug Drinen, founder of the Pro Football Reference website,8 AV9 is an attempt to measure an NFL player’s career impact on the league. Just counting “number of seasons as a starter” or “number of Pro Bowls attended” doesn’t capture a player’s weekly contribution to his team over the years. Whitman’s goal was to confirm the importance of generic athletic skills to an NFL career by comparing raw athleticism, as measured by pSPARQ, with success at the pro level, described by AV3 (the best three seasons of the player’s career).

While individual results will always vary from player to player, Whitman found an overall direct relationship of physical skills to production in the NFL when taken as a group. This satisfies two conditions of his maxim, that “very few poor athletes become good players” and that “most great players are great athletes.” It does not explain his third axiom: that “not all good athletes are good players.” Being physically elite in football may be necessary but not sufficient to being a game-changer.

Even with this relationship established and with the exploding emphasis on strength and conditioning in high school and college football, it would appear logical that average SPARQ combine scores should rise over time as players eventually get bigger, faster, and stronger. Surprisingly, a 2017 analysis10 found that the individual test results of the so-called offensive skill positions (quarterback, running back, tight end, and wide receiver) were relatively flat from 2000 to 2016. Average height and weight within each position has remained virtually the same. Forty-yard-dash times for quarterbacks and tight ends have stayed consistently at 4.8 seconds, while running backs and wide receivers have hovered around 4.6 seconds, +/−0.1 seconds, over the last seventeen years. Tests of quickness have become very tightly grouped across the four positions, while broad and vertical jumps have been virtually unchanged. Strength, as measured by the bench press, has declined for running backs, wide receivers, and tight ends, with quarterbacks rarely participating anymore.

Yet, the widely held assumption of forward progress in physical development remains. The data disagrees with claims by barstool pundits that athletes are bigger, faster, and stronger than the past generation.

Athleticism and sporting success are linked but not mutually exclusive in either direction. So, while Whitman showed that, overall, being a better athlete corresponds with being a better player, there is still an X factor to being a dynamic playmaker. Because physical training shows improvement in the form of numerical stats that can be tracked and compared, it is an easy target for training sessions. But we’re after those intangible improvements in knowledge, awareness, and decision-making that reveal themselves only subtly in competitive team situations. Coaches are convinced that they know a playmaker when they see one, but ideally they would prefer to create a training environment that proactively produces sport-related neural connections, just as speed training adds fast-twitch fibers and resistance training builds muscle.

Success Leaves Clues: Focusing on What Matters in the NBA Draft

With the first overall pick in the 1969 NBA Draft, the Milwaukee Bucks picked Lew Alcindor (thanks to winning a coin flip with the Phoenix Suns). Better known as Kareem Abdul-Jabbar, the three-time NCAA champion went on to become the NBA’s all-time leading scorer, a nineteen-time All-Star, and a Hall of Famer, and is still considered one of the greatest ever to play the game. The Bucks were able to win an NBA championship with him before he moved on to the Los Angeles Lakers, where he would win five more rings. Since that franchise-defining pick, the Bucks have had the luxury of three more number one overall picks. Kent Benson, Glenn Robinson, and Andrew Bogut were all meant to be linchpins for hanging more banners next to the lonely one from 1971. But lightning never struck again at the corner of 4th Street and Kilbourn Avenue in Milwaukee.

Hindsight does not require the prescription goggles that Abdul-Jabbar wore throughout his career. NBA general managers and coaches would give away lifetime season tickets to anyone with a crystal ball that predicts the career arc of a college or even high school player. Missing on the top pick can doom a franchise for years, not to mention the GM’s career. Since John Erickson, the first Bucks GM, made the obvious choice of Abdul-Jabbar back in ’69, the franchise has cycled through eight more GMs. Trying to extrapolate a star college player’s talent into a pro career of five, ten, or twenty years is as scientific as picking the draft lottery ping pong balls. Juggling the variables of individual contributions to a team sport goes beyond the basic stats of points, assists, and rebounds. Bottom line, does the team win more games because of this player? Can we allocate a percentage or share of a team’s wins to just one teammate? Bill James, the eminent sabermetrician, attempted this for baseball with a statistic he calls Win Shares, or “the number of wins a player contributed to his team.”11 Translating James’s original metric to basketball, Justin Kubatko, the creator of Basketball-Reference.com, derived win shares (WS) for basketball using similar logic as AV for football, by combining offensive and defensive statistics,12 to reflect a player’s overall contribution to a team effort.

Calculating a player’s career WS establishes their NBA legacy over time, while dividing it by minutes played, WS/48, reveals productivity even for a player plagued by injuries. Not surprisingly, Abdul-Jabbar holds the career record for total WS at 273.4, while Michael Jordan is first in career WS/48. Using WS/48, we can compare current players (with at least 15,000 minutes played or a little more than 312 full games) with retired legends to understand their emerging place in league history even while they are in mid-career. In the all-time top ten, active players Chris Paul, LeBron James, and Kevin Durant can stand with Hall of Famers like Abdul-Jabbar, Jordan, David Robinson, Wilt Chamberlain, and Magic Johnson to determine who is, literally, the most valuable.

Looking at the Bucks’ last three top picks, hindsight would tell them to look again. Benson, the big man out of Indiana, ended his twelve-year career with 33.6 WS, at a WS/48 rate of .103. However, it was Marques Johnson, the next Bucks pick after that, at number three, who would go on to be a five-time All-Star and have a much better career WS/48 at .162. Farther down the 1977 list were Jack Sikma, a seven-time All-Star with 112.5 WS, and Bernard King, a Hall of Fame inductee with 75 WS, at number seven and number eight, respectively. In fact, of the nine picks after Benson, all but one had at least the same or more WS.

Another poor season and luck in the draft lottery gave the Bucks another chance in 1994, when they selected Glenn Robinson out of Purdue with the number one overall pick. Despite a nine-year stint with the Bucks and two All-Star seasons, Robinson’s career (39.8 WS, .075 WS/48) paled in comparison to the second and third picks that year. Jason Kidd, (138.6 WS, .133 WS/48) would go on to be a ten-time All-Star over nineteen seasons. As the third draft pick that year, Grant Hill (99.9 WS, .128 WS/48), a seven-time All-Star, played eighteen seasons with four different teams.

Surprisingly, the lottery ping pong ball draw favored the Bucks again in 2005. Again, searching for a dominant big man, they made Andrew Bogut, a sophomore out of the University of Utah, the foundation of a rebuilding season. While Bogut’s career win shares (49.9 so far) are comparable to the second and third picks that year, Marvin Williams and Deron Williams, it was the fourth pick that, so far, has been the biggest miss. In the same twelve seasons, it is Chris Paul who has amassed 154 WS at an impressive rate of .250 WS/48, which is currently third all-time in NBA history. Already a nine-time All-Star, Paul trails only LeBron James and Dirk Nowitzki in career WS among active players.

In fact, in a 2016 Sports Illustrated analysis,13 the Bucks draft woes made several lists for “most regrettable” picks in NBA history. Benson was the ninth least productive number one pick, in career WS, from 1966 to 2007, while Robinson and Bogut were the eighth and ninth least productive number one picks compared to others available in the same draft from 1966 to 2016. Benson, Robinson, and Bogut may have filled positional needs at the time, but is there a better method that the Bucks coaching staff, or any coach, can use to identify future talent? Is there a signature statistic for “upside” or “untapped potential”?

Florida State University PhD students Jerad Moxley and Tyler Towne tried to nail down this elusive promise of skill growth in athletes jumping up to the next level. Being part of a legendary cognitive psychology department, led by Dr. K. Anders Ericsson and his cornerstone research on expertise, grounded Moxley and Towne in the fundamental differences between skill, talent, and performance. These terms are often muddied in the everyday world of sports coaching and recruiting, with early signs of skill being mistaken for long-term talent. Of course, performance is always relative to the level of current competition. The question on the minds of scouts and recruiters is: Will today’s skills that produce star performance grow into tomorrow’s talent potential at the next level of sport?

In 2014, Moxley and Towne set out to clarify these definitions so that they could tease apart the variables that NBA GMs should be focused on as they prepare for a draft. In their interpretation, skill is now, talent is future, and “untapped potential” is the difference. To even reach consideration for the NBA, players have already demonstrated quantifiable skill from their past performances. Different teammates, different opponents, and different coaches all contribute to an individual player’s résumé, leaving talent scouts to make the apples-versus-oranges comparisons. Solving for the variable of potential is the X factor.

“We will use the term ‘skill’ to refer to a player’s current level of performance,” Moxley and Towne explained.14 “We will simply define talent as different ceilings of performance. Talent can be conceptualized as containing two components, current skill level plus untapped potential.”

Has a player already peaked in college (or high school)? Has his growth curve already flattened out? Will the improved training practices, coaches, and facilities accelerate a player’s improvement rate after a transition? To give front-office management something to work with, Moxley and Towne contend that, “for scouts, it is important to note that a variable is only interesting if it is systematic and predictable by an observable metric.”15

Similar to the NFL Combine, the NBA also holds a pre-draft workout event where approximately the top sixty recruits are invited to participate. While there is no use of an overall summary statistic, like SPARQ for football, the players do complete five physical tests: two for agility, two for jumping, and one for sprint speed. Prior to 2013, a bench press tested for strength but was dropped in favor of the shuttle run. In addition, several anthropometric and physiological measurements are taken including height, weight, body fat percentage, hand size, arm wingspan, and standing reach. All data is available to the public at the NBA Draft website.16

Moxley and Towne believe that GMs and coaches use these eye-catching combine stats as tiebreakers between two players who are, in their assessment, otherwise equal. However, like an efficient stock market, they contend that a player’s physical attributes have already been factored into their past performances, so measuring them again at the combine—and using the data as a separate metric of untapped potential—is redundant.

Gathering data on all combine participants from 2001 to 2006, they built a model consisting of past college performance, quality of college program (as measured by number of players in the NBA), draft combine data, and performance in the first three NBA seasons to determine reliable predictors of success. For both college and NBA performance, win shares were used as the best available metric for individual contribution. The combine data was consolidated to player position (forward, center, or guard), the lane agility test, the no-step vertical leap, arm span, and weight.

Assessing the quantity and growth of NBA win shares in the first three seasons, the researchers confirmed their hypothesis, showing no significant contribution of anthropometric or athleticism variables, independent of college performance, to NBA success. In other words, the combine data added nothing to the analysis.

“The only variables that predicted NBA success were age, player’s college win shares and college quality,” Moxley and Towne concluded.

But that doesn’t stop scouts and GMs from relying on the combine data. In a secondary analysis, Moxley and Towne found that NBA draft order did significantly correlate with combine performance, despite the conclusion that this was not a reliable signal of untapped potential. While success at the previous level can be predictive of success at the next level, there’s no need to test players again on the physical building blocks. Traversing backward, then, we need to examine other possible influences on athlete development.

So the search goes on for the origin of the playmaker’s advantage, the stuff that consistently grows skill into maximum talent potential. To be sure, strength, speed, agility, and endurance are helpful if not mandatory. Still, in team sports that require integrated technical, tactical, emotional, as well as physical skills, there needs to be a locus of control. As you might expect, we think it is found above the neck.


WHAT IS A PLAYMAKER?

Peter Vint, PhD, former high-performance director of the United States Olympic Committee and former academy director of Everton Football Club:

Vision, decision-making, leadership, the ability to do the right thing at the right time, the ability to understand not only what’s happening in the game right now, but what is likely to happen next. Someone that from an academic standpoint may have [those skills], whether it’s an implicit or explicit understanding of situational probabilities, and leverage those in really effective ways at really important times.

Perhaps one of the easiest ways to describe [a playmaker] is that certain type of player who makes everybody around them better. I think that’s an easy phrase that a lot of coaches like to talk to their athletes about, but I think it’s the playmaker who is the one that ultimately ends up being perhaps a bit selfless . . . putting the team in the best position by doing what’s necessary at the right times.