Prospect rankings by the numbers
More and more, NBA teams are turning to analytics in the quest to maximize their return on coveted draft slots, or to unearth the gem farther down the board. When I talk to front-office types around the league, the topic of rating prospects almost always comes up. There is no magic in the process, either objectively or subjectively, and there certainly isn't any particular algorithm that's going to nail every prospect.
However, teams that ignore the analytical side of the game are working at a disadvantage. I've always liked the way Chicago Cubs president Theo Epstein describes it: The scouting side is one lens and the quantitative side is the other. It's only when you put both lenses together that the picture comes into focus.
I've been rating prospects for about five years using what I call the ATH system, which I've also used for various purposes on NBA-level topics as well. This year, I've made a few enhancements to the system (outlined in the chart at the bottom of this piece). The goal is to determine how a player's non-NBA production will translate to the game's highest level.
(Each player's prospect score (PROS) is denoted in parenthesis. Any score more than 100 denotes a deserving top-10 pick. Late first-rounders and early second-rounders -- always a gray area -- rate in the 65-75 range. Players fall off the draft board at about 50.)Top of the crop
1. Nerlens Noel, Kentucky (PROS: 124.2)
Noel is far from the slam-dunk top pick that the system suggested Anthony Davis was last year, but he does grade out as this year's top prospect. Noel's transcendent college blocks numbers help propel him to the board's top athletic rating, and he also rates as the best defensive rebounder. His ACL injury is factored in one respect: During his recovery, Noel has lost a lot of weight, coming in at 206 pounds at the Chicago scouting combine, which are the numbers I use when possible. This may have skewed how the system matches Noel's body type with past prospects.
2. Cody Zeller, Indiana (122.3)
From a per-possession-efficiency standpoint, Zeller ranks as the best prospect on the board. The excellent athletic markers he displayed at the combine hold up statistically, as he's got the third-best ATH rating in this class. Zeller lags in the skill area in terms of outside shooting and passing, but I'm still convinced Zeller is underrated in the prospect rankings I've seen.
3. Ben McLemore, Kansas (119.5)
Anyone who's seen McLemore stroke a 3-point shot will not be surprised that he recorded the top shooting score of the 134 prospects I measured, and he projects to put up the best 3-point percentage of any rookie next season. McLemore's upside is what's most intriguing to me: His projected five-year WARP is best in the draft, making him a solid option to go with as the top overall pick.
4. Alex Len, Maryland (118.3)
As another injured prospect, Len is a bit of a wild card, but he's got the second-best ATH on the board. Len's lower skill rating suggests that he is raw, but the system believes he'll be an efficient scorer and rebounder right off the bat.
5. Anthony Bennett, UNLV (116.7)
Bennett's athleticism shines through the projections, but the skill ratings lag a bit, probably because of his "tweener" body type. ATH is a little skeptical that he'll be an efficient 3-point shooter right away, and if it's wrong about that, look out. His Synergy isolation numbers aren't inspiring, which underscores the system's belief that Bennett won't be a volume scorer right away.
6. Steven Adams, Pittsburgh (114.6)
Adams is one of the big risers on the board in terms of disconnect between metrics and scouting rankings. He's got the best-projected rookie rates in 2-point percentage, overall rebound percentage and block percentage. Yes, his block projection is even a hair better than Noel's. Body type has a lot to do with that. Consider Anthony Davis, who projected to lap the field in rookie block rate last year. He indeed was an excellent shot-blocker, but Andre Drummond and Festus Ezeli both put up better rates. Despite a foul rate that looks problematic, Adams is a worthy project.
7. Trey Burke, Michigan (110.1)
One of the draft's most vexing questions is who is the top point guard prospect? ATH likes Burke, though not by enough to settle the issue. The system sees him as a bit of a ball dominator, with rates in usage, assists and turnovers, and has severe concerns about his defensive prowess.
8. Michael Carter-Williams, Syracuse (107.3)
Carter-Williams has the edge on Burke on the defensive end, where Synergy gave him the third-best score against isolations of any player in the draft.
9. Otto Porter, Georgetown (107.0)
Porter is more well-regarded by the scouts than the stats, though a No. 9 ranking is nothing to sneeze at. However, his five-year WARP projection ranks just 12th in this draft class.
10. Victor Oladipo, Indiana (105.0)
Scouts love Oladipo's defensive upside, but it just doesn't translate statistically. ATH rates him 70th in defensive rating, 41st in steals and 52nd in blocks, and Synergy ranks him 81st against isolations.
11. C.J. McCollum, Lehigh (99.
McCollum may be the top scorer among next year's rookies because of his combination of skill and volume. His traits are heavily balanced toward the skill over the athletic, which is a concern. Also, his No. 13 ranking in five-year WARP puts him behind Burke, Carter-Williams and a soon-to-be-named point guard at his position.
12. Dennis Schroeder, Germany (98.
And that unnamed point guard is Schroeder, though as a player who is largely unproven in high-level leagues, he's a bit of an enigma. That said, his upside is tantalizing, with a five-year projection that is the top at his position.
13. Sergey Karasev, Russia (96.6)
Everything you've heard about Karasev's pure shooting ability shines through statistically, and he rates as the most-efficient offensive player in this draft. He also rates 130th out of 134 defensively, and his athletic rating is 133rd, so his strengths -- and weaknesses -- are apparent.
14. Jeff Withey, Kansas (90.7)
Withey is another big man whose lean body type is holding back his projected rebounding and block rates, and his age (23) is a drag on his five-year projection. His rookie-season projection is promising enough that he could be a rotation center immediately on the right team.
15. Gorgui Dieng, Louisville (87.3)
If healthy, Dieng could easily emerge as the most NBA-ready defender of the draft. He's the same age as Withey, though the system doesn't know about his overall lack of basketball experience. The passing abilities that the scouts like so much don't shine through the translations for Dieng, either. They'll need to, because he's not going to be a scorer.
16. Dario Saric, Croatia (84.2)
17. Glen Rice Jr., Georgia Tech (83.1)
18. Kelly Olynyk, Gonzaga (82.7)
19. Mason Plumlee, Duke (78.
20. Jamaal Franklin, San Diego State (78.5)
21. Jackie Carmichael, Illinois State (76.1)
22. Tony Mitchell, North Texas (75.1)
23. Mike Muscala, Bucknell (73.4)
24. Andre Roberson, Colorado (72.9)
25. Laurence Bowers, Missouri (71.7)
26. Allen Crabbe, California (71.7)
27. Rodney Williams, Minnesota (70.7)
28. Shabazz Muhammad, UCLA (69.
29. D.J. Stephens, Memphis (69.5)
30. Rudy Gobert, France (69.5)
In the last half of the first round, you'll notice that Gobert's statistical profile is underwhelming. There are a few unexpected names, most noticeably Bowers and Williams, but given that the system is built on a foundation of athletic markers, their inclusion might not be so surprising. If Bowers can further hone his face-up shot and develop some range, I love him as a defensive specialist, while Williams reminds me at times of Tony Allen. The system spit out the following about Williams: His most likely career path is that of Gerald Wallace. Neither Bowers nor Williams is a cinch to be drafted at all, much less in the first round. Meanwhile, Muhammad is the latest Ben Howland player whose statistical translations are highly pessimistic. Also, you can see where Saric rates if he does another about-face and stays in the draft.
32. Deshaun Thomas, Ohio State (68.1)
38. Kentavious Caldwell-Pope, Georgia (64.4)
41. Nate Wolters, South Dakota State (64.0)
50. Shane Larkin, Miami (Fla.) (56.3)
Caldwell-Pope and Larkin are the highly touted prospects who fall the farthest in the ATH system. Caldwell-Pope has good stand-still shooting indicators but is held back by a poor defensive translation that conflicts with scouting reports. Larkin has the top projected assist rate in the draft, but suffers from poor shooting efficiency and the second-worst turnover rate on the board. The system undersells both of these players, but it does raise some red flags in their performance record. Thomas and Wolters merit special mentions because they top the board in the Synergy stats. Thomas has the best defensive isolation score, while Wolters tops the offensive end.
ATH system enhancements
1. Each player's ATH rating, or athletic rating, is calculated for his numbers in college, overseas or the D-League. It's based on height-adjusted measurements for rebounding, foul drawing, blocks and steals. This metric largely determines how much of a player's non-NBA production will be retained.
2. A skill rating, or SKL, is also calculated. It's based on the more gentle parts of the game, such as free throw shooting, 3-point shooting, turnover rate and assist rate, all of which are also adjusted for body type.
3. Other measurements are recorded for the purposes of player classification: height, weight, likely base position, age, a scouting rating based on combine measurements. I also enter the scout-based prospect rankings from our own Chad Ford to use for overseas players who haven't played in a high-level league, or for American players who didn't play college ball, such as Enes Kanter. In such cases, there just isn't objective data to go on.
4. To refine projections for players like Glen Rice Jr., who play in the D-League prior to entering the draft, I've added numbers from that circuit and blended them with their college performance. The bigger the sample, the better.
5. Finally, I've taken advantage of the terrific data from Synergy Sports Technologies to help refine player classification. I may add more play types in the future, but for this year, I was most interested in how the players rate in isolation situations, on both offense and defense.
The ATH metric is still the primary component in translating the numbers forward, but the other traits are used to better match a player's baseline projection with the historical database, which in turn suggests the probable career path for the player. The system now spits out extremely conservative five-year projections, which aren't meant to pin down a player's exact future value, but more to suggest the likely direction of his growth curve beyond rookie-year projections.