I always enjoy thinking about player comps. I’ve included them in previous scouting entries, like my posts on underrated and overrated draft prospects and summer league players to watch. For prospects and young pros, a comp is really a form of projection that tends to focus on style more than value. The central question is this: which current or former player(s) will the young player resemble, or appear similar to, at the NBA level?
Comps are often stated in terms of a player’s “upside” or “ceiling” (optimistic projection) and “downside” or “floor” (pessimistic projection), though I’d argue that the term “floor” is a misnomer as a general rule. Typically a prospect’s stated floor is someone who cracked an NBA rotation for several years, maybe even as a starter, and thus it’s far from a worst-case scenario.
For a particularly egregious example, I’ll never forget an ESPN writer listing Horace Grant—starting Power Forward on 4 title teams, All-Star, and Top 100 player all-time by quite a few statistical measures—as Noah Vonleh’s floor. With all due respect to Noah Vonleh, if Horace Grant is actually any prospect’s floor, that prospect ought to be the consensus #1 pick in the draft.
Along these lines, I’d consider Joe Johnson a natural comp for Josh Jackson—a playmaking, 6’8” freshman wing at Kansas who excels in isolation and transition—and it’s reasonable to project that Jackson will be a more disruptive defender, but it would be absurd to call Joe Johnson his floor. This is true despite both DraftExpress and nbadraft.net listing Jackson as a top-3 pick in their current mock drafts, with the latter comparing him to Kawhi Leonard. An example from Kawhi’s own draft class illustrates why. DraftExpress said that the #2 pick in that draft’s “worst case” was a “More Athletic Antawn Jamison.” Now in his sixth year in the league, Derrick Williams doesn’t have a single season that measures up against Jamison’s decade-long prime.
Joe Johnson, though never an all-around difference maker on Kawhi’s level, was the best offensive player on a few 50ish win teams, a 7x All-Star, and a perennial above-average starter for over a decade. To say that Joe Johnson is Josh Jackson’s NBA floor would be nearly as crazy as saying that Jack Johnson is his floor as a boxer, (Shoeless) Joe Jackson is his floor as a hitter, Josh Johnson is his floor as a pitcher, or Joshua Jackson is his floor as an actor playing a hockey player. It simply would be a massive overbid considering how few prospects reach that level.
In order to identify more meaningful floors, ceilings, and realistic projections, similarity scores can be especially helpful. Rather than relying solely on your impressions and recollections of players, you can start with a statistical model that spits out a list of relatively comparable players to consider. Check out fivethirtyeight.com’s CARMELO projections for a great example. The CARMELO projection system uses a database of player seasons since the 1976 ABA-NBA merger and determines comps based on various stats as well as other measurable characteristics like age, height, weight, and draft position.
The max similarity score is 100 under the CARMELO model, and points are subtracted for statistical differences in 19 categories. Returning to the Noah Vonleh example, CARMELO identifies his nearest comp as Darrell Arthur, with a similarity score of 57. We can identify optimistic and pessimistic projections, or ceilings and floors, by looking at the other names on the list. Al-Farouq Aminu and Wilson Chandler would seem to be particularly favorable outcomes, whereas Yi Jianlian and Johan Petro are lower-end comps.
Of course this isn’t an entirely fair exercise, given that CARMELO has 2 years of NBA performance on which to evaluate Vonleh and the ESPN writer had only a single year of Indiana Hoosierdom to go on. To even things out, we can look to this year’s #8 pick Marquese Chriss. CARMELO actually lists Vonleh as his nearest comp, with a similarity score of 55. Based on his other comps, his ceiling could be as high as Chris Bosh, though Steven Hunter rather than Horace Grant would seem to be his logical floor. [According to CARMELO it’s ‘Steve’ Hunter, yet I’ve never heard him called Steve before. Leading up to the draft I identified him as a comp for another prospect, Skal Labissiere.]
I formulated my own, much less comprehensive similarity scores model using a database of all player seasons from 1978 to 2015. For each statistical variable that I selected, I compared players as follows: ((Player X – Player Y)/ League Avg)^2. In other words, I started with the difference between two players in a given stat, then I divided that number by the league average to put each stat on an equal footing, and the result is squared so that every difference is a positive number. Then my overall similarity score is the summation of those numbers, with lower scores indicating a closer relationship.
Using the variables ORB%, DRB%, AST%, TOV%, STL%, BLK%, 3PAr, eFG%, FgUsg%, ORtg, and Points Produced—a Dean Oliver stat representing total offense, which combines scoring, assists, and offensive rebounds—I ran comps for a few players from the 2013 draft. The following comps are based solely on the 2014-15 season [the last season for which I have data]:
Most of these results seem to make sense stylistically. Oladipo’s list is filled with medium to high usage combo guards, though it includes both efficient and inefficient scorers. Oladipo’s unremarkable shooting efficiency produces player comps that are more varied in that area, while similar in others. Jeff Hornacek is not only his top comp but also the only player appearing multiple times.
On the subject of Utah Jazz standouts, Rudy Gobert naturally is similar to a bunch of imposing rim protectors, most favorably 3 seasons of prime Dikembe Mutombo. Looking beyond this Top 10 List, there are 5 more Mutombo entries among Gobert’s Top 25 comps by similarity score, such that Mutombo accounts for about one-third of the Stifle Tower’s most comparable player seasons.
Both Oladipo and Gobert signed 4-year extensions recently, Oladipo for $84M and Gobert for $102M. If their careers progress like Hornacek and Mutombo, those deals will quickly look like bargains.
Now to a couple players who didn’t reach extensions:
Caldwell-Pope’s comps are all jump-shooting wings who typically contributed more offensively than defensively. Whether these players were effective starters or not hinged on their defensive ability, with gritty defenders like Wes Matthews and Caron Butler playing major minutes on playoff teams and sieves like Marcus Thornton relegated to a situational bench role. I’d place KCP firmly on the Wes Matthews side of the spectrum, so in my estimation he’ll merit a big contract this summer.
Nerlens Noel’s profile is a bit trickier. His 2015 season is reminiscent of late-career Hakeem Olajuwon, when the Dream scaled back his minutes and usage considerably. Hakeem was at least 35 in each of these similar seasons, and none of his earlier years makes Noel’s top 1,000 comps. Noel also has the rookie seasons of Anthony Davis and Kevin Garnett on his list, plus Tyrus Thomas—basically a collection of ridiculously athletic young big men with defensive production initially exceeding their offensive skills.
Two more Tyrus Thomas seasons crack the Top 15, whereas KG and Anthony Davis disappear based on their rapid improvement. Far from rapidly ascending the ranks of NBA big men, Noel followed up his 2015 season with nearly an exact replica last year. His next most common comp, with 4 appearances in the 11-30 range, is Terry Tyler—an athletic, defensive-minded SF/PF from the 80s who averaged 10.2 points, 5.4 rebounds, and 1.5 blocks in 25 minutes per game. In sum, Noel’s comps are all over the place in terms of career performance, from role players to superstars, and I’m still not sure what to make of his potential or likely projection based on this list.
As much as I like similarity scores, I find that Play Index tools on the Sports Reference sites have lessened their importance. Once you identify a player’s core strengths and weaknesses statistically, you can go to the Player Season Finder on Basketball-Reference.com or Sports-Reference.com/CBB, plug in those statistical attributes, and find a list of comps in one click. For example, with respect to Nerlens Noel, I looked on Basketball Reference for players 6’9” & up with per-36 averages <18 points, >1.5 steals, and >1.5 blocks, and with <55% fg and <1 3PA/G. Noel’s 2015 season ranks #25 by Win Shares on the resulting list, which includes a couple of those old-Hakeem seasons, Len Elmore, and other top comps Tyrus Thomas and Oliver Miller, all within several spots of Noel.
This Play Index search shows that Ben Wallace may be a better reflection of Noel’s upside than Olajuwon or Garnett, though the presence of several early-20s Tyrus Thomas seasons bracketing Noel’s past 2 years is concerning. Much like Noel, Thomas entered the league as a very good defender based on his athleticism, size, and energy, while contributing little offensively. Thomas never figured out that other end of the court, and once he lost some of his athleticism due to injury he found himself out of the league entirely. Any team that commits itself to Noel this summer with a long-term deal will need to hope his path is very different. If ultimately Noel’s career ends up closer to Ben Wallace than to Tyrus Thomas he’ll have done quite well.
With this alternative method at our disposal, let’s return to Josh Jackson. Play Index tools for college players are somewhat more limited, with fewer filters and most advanced stats only available for the past handful of seasons. Thus, the first step in finding comps is identifying Jackson’s core attributes by basic stats. Throughout the year Jackson has averaged 15 points, 6 rebounds, 3 assists, 1.5 steals, and more than a block a game. He’s also been doing this as a freshman, albeit an older one set to turn 20 in February. Searching Freshmen and Sophomores with per-game averages of 13+ points, 2.5+ assists, 1.3+ steals, and 0.8+ blocks yields the following result:
Nine of those players have been drafted: Ron Artest, DeAndre Bembry, Francisco Garcia, Paul George, Otto Porter, Ben Simmons, Evan Turner, Dwyane Wade, and Bonzi Wells. In my opinion Artest isn’t an ideal comp because he was built like a tank whereas Jackson is much slimmer and quicker off the dribble. Bembry and Simmons aren’t useful comps given their limited to non-existent NBA experience, and Simmons would’ve been eliminated anyway if I had more filters to add an upper limit on rebounding.
That leaves Wade and Paul George as high-end comps, Francisco Garcia and Evan Turner as low-end comps, and Otto Porter and Bonzi Wells somewhere in between. In terms of an expected outcome (or 50th percentile projection) for Josh Jackson, based on these comps I’d think along the lines of a motivated and well-mannered Bonzi Wells. Bonzi averaged 17.7 points, 6.3 rebounds, 3.4 assists, and 1.8 steals per 36, with an above-average 17.3 PER, .109 WS/48, and 1.5 BPM, during his first several seasons in Portland before off-the-court issues and poor conditioning derailed his career.