Sports

How Analytics Can Help Identify NBA Draft Steals and Busts

"Moneyball"-style analytics have come to the NBA Draft, and while they remain a work-in-progress, they can provide valuable insight into which players to choose and avoid. Here's what you need to know.

by Kevin Broom
Jun 23 2016, 4:30pm

Derick E. Hingle-USA TODAY Sports

The NBA Draft is truly a lottery, and not just in how the first few picks are awarded. Teams are making multi-million dollar bets on players—often teenagers—whom they hope will one day become high-quality pros. The hurdles and pitfalls are numerous, but the potential rewards are huge.

Teams have invested in an array of approaches for evaluating draft prospects, including old-fashioned scouting; extensive background checks; personal workouts and interviews; medical exams; drills and exercises intended to predict susceptibility to injury; and yes, statistical analysis.

While there are an array of superb statistical approaches available for analyzing the contributions of players already in the NBA, the tools for evaluating prospects are less robust, despite the efforts of several smart independent analysts and the first-rate minds employed by actual teams.

Read More: Meet The Sports Psychologist Training The Minds Of The NBA's Top Draft Prospects

Indeed, the effort to inject some statistical science into the inexact art of choosing players hasn't prevented teams from mis-evaluating prospects. As this article from Nylon Calculus shows, it sometimes takes a while—even until the second round—for the league to find the best players in a draft. "Moneyball"-style analytics have come to the NBA Draft, but they're very much a work in progress.

Still, thoughtful consideration of player measurables—box score stats, physical attributes, age, and how those compare to others—can help any team avoid busts, unearth overlooked gems, and provide a much-needed corrective to the sometimes-deceiving eye test.

For example, most statistical analysts (myself included) thought Jae Crowder and Draymond Green should have been highly-rated prospects in the 2012 draft. The eye test disagreed. In a conversation shortly before the draft, one team executive told me, "Jae rates high in our statistical models, and he's a great kid, but he doesn't have a position."

TFW it turns out you have an NBA position, after all. Photo by Jason Getz-USA TODAY Sports

Oops. The numbers turned out to be right. While many teams and analysts have their own unique formulas for evaluating players, there are four key statistical categories to keep in mind as you watch tonight's draft:

Blocks, steals and two-point field goal percentage. These are all markers for applied athleticism. Guards need a two-point percentage above 50 percent. Bigs need to be close to 60 percent. For a sleeper pick, look at a guy like Stony Brook's Jameel Warney, a power forward who shot 63 percent on twos, and tallied 4.8 steals and blocks (combined) per 40 minutes. The lack of steals and blocks are potential red flags on likely lottery pick Jaylen Brown.

Overall offensive efficiency. An offensive rating below 110 points produced per 100 possessions is a red flag, regardless of position. This is a possible warning sign on this year's reputed top point guard, Kris Dunn. He does some things well, but very high turnover numbers suggest it might be smart to pump the brakes.

All-around production. Big men with high assist numbers tend to be valuable NBA players. Big man assists are also often an indicator of defensive potential. Similarly, guards who rebound well often transition well to the NBA. The best passing big this year: perspective No. 1 overall pick Ben Simmons, who averaged 5.5 assists per 40 minutes. For an unsung prospect, check out Stephen F. Austin's Thomas Walkup. He dominated for three years, doing a lot of everything with extreme efficiency. He's likely to go undrafted, but rates like a mid-first rounder.

Free throw shooting. When evaluating shooting potential, the tendency is to look at three-point percentage. This isn't necessarily wrong, but sample sizes are often quite small in college and international hoops. Look instead at free throw percentage to get a better read on a player's shooting prowess. Kentucky's Tyler Ulis shot an unimpressive 34 percent on threes this year, but his 86 percent free throw shooting suggests his jumper won't hold him back in the NBA.

TFW the numbers love you, because you're Thomas MF'ing Walkup. Photo by Robert Deutsch-USA TODAY Sports

Keep in mind: just knowing which numbers matter the most isn't enough. Not on its own. Any reasonably predictive analysis also has to place those numbers into an appropriate and useful context.

As someone who has studied both draft history and player analysis—and worked on my own performance evaluation system—I think there are 10 keys to using stats to evaluate NBA prospects:

1. Define the question. This seems simple, but it's critical to decide exactly what kind of information team decision-makers want. Who's the "best" player might shift depending on what the team is trying to find. Does the general manager want to know who's going to contribute immediately? Are they looking for the player who's going to have the best four-year value (the period of complete team contract control for first round picks)? Do they want to know who is going to have the best overall career? Some analysts even estimate the odds that players will achieve certain performance levels (All-Star, starter, journeyman, rotation player ... or bust).

2. College and international stats aren't one-size fits all. Quick, who was the leading scorer in college basketball this season? The correct answer: James Daniel, a 5-foot-10 guard for the 12-20 Howard Bison. He's not in the draft this year, and likely wouldn't be selected if he was. The point is that level of competition matters. Two players with identical stats can have markedly different pro potential depending the opposition they faced. As an NBA executive once said: "You have to keep in mind that in college or international ball, there's maybe one or two NBA [quality] players on the floor at a time."

3. Don't dismiss top performers from bad teams or who produced against lesser competition. There aren't many guys who faced weak competition or played for a bad college team who went on to outstanding careers in the NBA. But, those who made that leap have at least one thing in common: they dominated. Paul Millsap, Paul George, Kenneth Faried, and Damian Lillard spring to mind as prime examples. In this year's group, look to Walkup, who shot 60 percent from the floor and 82 percent from the free throw line while rebounding like a power forward, assisting like a point guard, swiping the ball like a shooting guard, and avoiding turnovers like a low-usage role player. He's likely to go undrafted and become a free agent bargain for the right team.

4. Don't ignore production. Team decision-makers continue to pick terrific athletes under the theory that those guys could be great someday if they just develop one or more basketball skills. What they seem to overlook is the poor record of similar players developing those skills. As Lamar Odom once said of Javale McGee: "The game is basketball, not run and jump." While some are able to identify relevant skills through careful scouting, there's much to be learned from careful analysis of box score stats. Eye-popping athleticism is great, but what matters is application of those physical tools—meaningful athleticism shows up in the numbers. In recent years, Tony Wroten and Perry Jones fit the McGree profile. This brings me back to Jaylen Brown, whose stats at Cal were remarkably pedestrian ... except for his turnovers. Which were shockingly bad.

5. Look at relevant physical attributes. Height doesn't mean anything. What matters is standing reach and wingspan. Millsap, for example, was "undersized" coming out of Louisiana Tech. But his standing reach was about average for an NBA power forward. This year's version, albeit not to the same degree as Millsap, might be Stony Brook's Warney, who has adequate size and outstanding production.

6. Age matters. Older players should post better numbers because they're older, more experienced and more physically developed than the typical freshman. So, if you have similar production against similar level of competition, pick the younger player. Take Kentucky's Jamal Murray and Providence's Dunn. Superficially, their numbers look somewhat similar, and some might rate Dunn higher. But Murray produced as a freshman, and played nearly double the number of minutes Dunn did in his freshman year. On the other hand ...

Expected NBA Draft lottery picks Jamal Murray (left) and Brandon Ingram were productive as college freshmen. Photo by Dennis Wierzbicki-USA TODAY Sports

7. Don't over-value youth. Pick a lane, right? I know I just said to pick the younger guy, but that's when most things are about equal. Often, those highly productive seniors will become productive pros.

8. Beware the lure. These are guys who look like good players, move like good players, seem to have skills like good players, but are not good players. Think the aforementioned Perry Jones, Tim Thomas, Andray Blatche or Kwame Brown. There are players like this in every draft. This year, Brown seems a to fit this mold. Scouts love him, and he has an impressive, NBA-ready body. But he actually doesn't produce very much when he's on the floor.

9. Pick the oddballs. If you can select the next Anthony Davis or LeBron James or Shaquille O'Neal, then of course, you do it. No questions asked, and no real need for deep analysis. Not when a player is a physical prototype and super-productive. Of course, few teams have that opportunity. They're left choosing from what's available. NBA history is replete with players chosen late because of cosmetic defects like height, weight, weird body shapes or injury histories. Thing is: many of those players were highly productive in college, but were dismissed by pro scouts because they "knew" those guys couldn't make The Leap to the NBA. The Boston Celtics won 48 games this year with "defective" players like Isaiah Thomas (too short), Jae Crowder (too short, sensing a theme), and Jared Sullinger (too fat). The Spurs got productive years from DeJuan Blair, a highly-productive college starter who they got in the second round because he didn't have an ACL in either knee, because he was "too short" to play power forward, and because he was built like an offensive lineman. A list of productive players deemed "too short" in the draft would stretch into the ether, but the point holds: guys who were productive against good competition at lower levels tend to be productive when they reach the NBA, even if they seem to lack ideal physical attributes for the NBA.

10. Don't worry too much about fit. The current thinking—perhaps it has always been like this—is The Puzzle Theory. In this construction, teams add certain packages of skills to the roster, and the combination of skill packets make the team more effective collectively. Support for this approach is typically found by pointing to whichever team just won a title and marveling at how well everyone "fits." Whether this is the right approach or not to roster construction—I prefer the simpler approach of acquiring the most productive players possible while accounting for each of the different jobs that need to get done—it's definitely the wrong approach to the draft. With few exceptions, GMs don't know which prospects will become good professionals and which will be also-rans. If decision-makers can't reliably pick who the best player will be, adding "fit" into the mix is a bridge too far. Just pick the best player available and be done with it. Good players usually find a way to work together. If they can't, someone can always be traded.

Want to read more stories like this from VICE Sports? Subscribe to our daily newsletter.