During the run-up to tonight's NBA Draft, it seems like the idea of rating prospects statistically has gained more of a foothold than ever before. That observation pertains more to the coverage of the game than the machinations of team insiders, but it's nice to see analysis continue to see the light of day. During the chat I conducted a couple of days ago, the player I was asked about more than any other was North Carolina's Ty Lawson. That was the result of John Hollinger's system flagging Lawson as the top-rated prospect in this year's draft class.
That apparently provoked a lot of response, as evidenced by this bit from Chad Ford:
"I'm not an expert on stats, but I can tell you that for most of the more sophisticated front offices in the league, statistical analysis plays a very significant role in the draft process. Teams like the Rockets, Mavericks, Celtics, Thunder, Spurs, Sixers, Nuggets (I could go on and on) may use slightly different formulas and methodologies, but all of them are using statistical models to analyze the draft. For a handful of teams, what they find is a major factor in whom they choose to draft."
At the other end of the spectrum from Lawson, these models have also identified USC's DeMar DeRozan as a potential draft bust. His name came up in the Basketball Prospectus roundtable discussion we conducted earlier this week. The esteemed John Gasaway isn't a fan; neither is my system. Kevin Pelton's translations agreed about Lawson, though they didn't have him as the top overall prospect. Those same numbers ranked DeRozan as the #2 shooting guard prospect.
In the aforementioned chat, I also mentioned my translations a couple of times. My system has been built up around my efforts to quantify the athleticism of a player. This stems from some research I did with NBA numbers a couple of months back, which you will be reading more about in the future. My new metric, ATH, appears to have uses for identifying things like whether or not a player can transition to higher-usage roles and when he's hit a point in his career when you can expect his efficiency and/or usage to crater.
When I applied the same methodology to the college numbers of every player drafted from 1997 through last season who had played at least 100 NBA games, I found more possible uses for ATH. First, it seems to help identify, at least to a certain extent, how much a player's college productivity is going to be retained in his transition to the pro game. Also, ATH can be used to flag players whose efficiency and/or usage may have been wanting at the college level, but who still have ceilings that might allow them to become useful NBA players. Conversely, the metric can (maybe) help identify the DeRozans of the world, who just don't seem to have the tools to excel in an NBA environment. Finally, I can see a time when I can use ATH to derive confidence intervals in a player's translated college stats.
With those notions in mind, I'm going to throw out some sleeper prospects for this year's draft. The system rates players based on height-adjusted offensive efficiency which is adjusted for usage rate, ATH and a defensive factor derived from the player's height-adjusted rate of blocks plus steals. The numbers are also adjusted for the quality of the competition a player faced, whether it be in college or in Europe. At this point, however, I did not adjust for age. That's certainly the next step.
The ATH component, the product of my effort to quantify athleticism, is derived from the following height-adjusted measurements: rebound rate, steal rate, block rate and foul-drawing rate. As you can see, there is nothing in the calculation that acknowledges field-goal attempts, yet I've found that it's a good metric for predicting usage. (Or at least the ability to maintain or increase usage and efficiency when moving between contexts.) I should also add that when I refer to athleticism, I'm speaking more of what you might call applied athleticism. The formula doesn't know how fast a players is or how high he can jump. It instead looks at how a player applies these traits and converts them into basketball productivity. So a player, like DeRozan, who has superior physical traits to, say, Tyler Hansbrough, can rate much lower in ATH. That can be as much evidence of passivity or aggressiveness as anything else.
Most of you reading this site probably have at least a general idea of correlation. That is when you compare two groups of numbers and measure how well they correlate with each other. A perfect correlation is positive 1. Negative 1 means that the numbers have perfectly opposite correlation. Zero means no correlation at all. When you're testing theories using this method, you're hoping for a score as close to 1.0 as you can get.
When I ran the NBA numbers for my set of draftees against their college numbers, I found the following correlation scores:
Offensive rating: 0.26
Usage rate: 0.17
So ATH is a metric that translates very well from one level to another. That, in itself, is useful knowledge, but enough about methodology. The draft is just hours away now, so let's see who this cockamamie rating system thinks is getting the short shrift on draft boards.
- DeJuan Blair, Pittsburgh: Blair is to my system was Lawson is to Hollinger's. Yes, that means he's my top-rated prospect. Don't get me wrong, I'm not advocating that Clippers grab Blair rather than Blake Griffin (who ranks just behind Blair and Stephen Curry in my system), but I do think whichever team drafts Blair is getting one of the draft's top players.
The knock against Blair is his height. He's generally listed at 6'7", but he's more likely pushing 6'6" and plays the four-position. However, he's got a 7'3" wingspan and that trait seems to be reflected in his numbers. The system awards Blair with the draft's highest ATH rating, foresees a player with good efficiency to go with around 24% usage and an excellent defensive factor. I should acknowledge that because I'm working with height-adjusted numbers, there is a bit of a bias towards undersized rebounding machines, like Charles Barkley and, to a lesser extent, Jerome Lane and Danny Fortson. If Blair turns out to be more similar to Sir Charles than the latter pair, that will be quite a boost to my new system's credibility.
- Leo Lyons, Missouri: Another possible systemic pitfall here is that players in Mike Anderson's frenetic system may end up with inflated ATH indicators. Or that may really be excellent athletes. On that, I can vouch for Lyons, who projects to be a high-usage, nominally efficient NBA player with just above-average defensive abilities. However, Lyons' ATH rating fingers him as having a high ceiling and more of a sure bet to translate his game to the NBA level than some of the more highly-touted prospects. Lyons isn't a sure bet to be drafted, so I hope some team takes a flier on him in the second round.
- Lester Hudson, Tennessee-Martin: Yes, he's old. Nearly 24. Nevertheless, my system thinks that Hudson can be a very efficient NBA offensive player with 25% usage and very good defensive indicators. His ATH rating is good, but not off the charts, so that diminishes my confidence a bit, as does Hudson's age, which isn't accounted for in this system. However, I could see him emerging as a Louis Williams type.
- Ty Lawson, North Carolina: My system loves Lawson, too. From what I've been reading about the individual workouts, Lawson may be moving up the draft board. We'll see. The ATH system loves his offensive efficiency and thinks his athleticism will enable Lawson to translate that to the pros.
- Taj Gibson, USC: Gibson is a guy who has emerged as a possible late first-rounder. My system likes him, primarily for his defensive skills, which mark him as one of the top five defenders in this class. His ATH rating is only a little above average, so buyer beware.
- Jermaine Taylor, Central Florida: Taylor projects to be a 29% usage player with fair efficiency who lacks defensive skills. I'd be completely skeptical about that usage rate if not for a very good ATH score.
- Danny Green, North Carolina: Green's ATH rating is below average, which concerns me. I also see his offensive game being limited to a Shane Battier-like existence. However, my system also see Green as being a possible Battier-like defender. The only perimeter defender in this draft I like better than Green is Tyreke Evans.
- DeMarre Carroll, Missouri: Again, this may be a bias towards Mizzou's system, but my system sees Carroll as a solid athlete and a potentially disruptive defender.
- Marcus Thornton, LSU: The ATH system sees Thornton as an efficient scorer in the 25% usage range with an ATH rating high enough to make me believe those observations are real. His defense doesn't project as well.
- Jerel McNeal, Marquette: I'm not sure McNeal stands out a whole lot in people's minds from fellow Marquette prospects Wesley Matthews and Dominic James, but he should. I see him as being a first-round talent that will be available in the second round, while the latter pair aren't on my board at all. McNeal doesn't have one standout trait in my system, but he's solid across the board and sports an ATH rating that makes me think he's got real NBA upside.
Bradford Doolittle is an author of Basketball Prospectus.
You can contact Bradford by clicking here or click here to see Bradford's other articles.