The following is from our resident Raptor blogger Devin Dignam of NBeh?. I’ll turn the floor over to him.
I’ll offer more in-depth coverage shortly, but here are my quick thoughts on the 2011 NBA draft.
Importantly, my evaluations rely on Position Adjusted Win Score (PAWS). PAWS takes Win Score per 40 and adjusts it by position, because each position has different average values.
To calculate this, take the WS/40 of a player, subtract the average positional value, and then add the average value for all players. The average values that I’m using are the same ones used by David Berri, who is one of the brains behind Win Score and Wins Produced. With players who are listed as playing two positions, I use a simple average of the two positional values in question. For those who are curious, the values are as follows:
- PG: 7.4
- SG: 8.4
- SF: 9.95
- PF: 12.59
- C: 12.32
- All players: 10.17
Ranking all players by PAWS gives us a good idea of how well a player performed in their various leagues. A word of warning, though: NCAA PAWS/40 does not correlate perfectly to NBA success, and Euroleague PAWS/40 is even worse. There are players with historically good NCAA PAWS (like Michael Beasley) who don’t turn out to be good players, just as there are players with poor or mediocre NCAA PAWS who turn out to be pretty decent players (John Wall and Derrick Rose come to mind). For the most part, though, players with a PAWS of 12 or higher usually end up being good NBA players, and players with a PAWS of under 7 end up being below average players. Euroleague PAWS are far less reliable due to the way basketball is played overseas, and to my knowledge, no one has examined how well Euroleague PAWS correlate with NBA productivity. Keep this information in mind while using this data.
1. Who won the draft?
Without question, the Denver Nuggets made the best choices on draft night. With the #22 pick, the Nuggets managed to get the NCAA player who posted the highest PAWS/40 (17.28) in the draft, Kenneth Faried. Then they picked up Jordan Hamilton at #26, who posted the 21st best PAWS/40 (at 10.72) in the draft. Of course, the cost of acquiring the #26 was swapping Raymond Felton for Andre Miller, but (ignoring age) that was largely a lateral move, as Felton and Miller produce at about the same rate and are a very close match salary-wise. With a lineup of Ty Lawson, Arron Afflalo, Jordan Hamilton, Faried, Nene, and Andre Miller, Wilson Chandler, Danilo Gallinari, and Chris Andersen coming off the bench (or in some such combination), Denver looks really scary.
I bet they miss us now
2. Who lost the draft?
With the #6, #18, and #34 picks, the Wizards had a chance at a huge draft haul. Instead, they walked away with Jan Vesely, Chris Singleton, and Shelvin Mack. Vesely posted a PAWS/40 of 6.76 overseas, which was only good enough for 66th place among all the players who were drafted. Singleton (8.50, 47th) and Mack (8.67, 43rd) weren’t much better, but at least they were somewhat close to average (PAWS/40 of 10) and not below the cutoff line (PAWS/40 of 7).
3. Who made the biggest steal?
There were a lot of second round steals this year, but to me, the team that had no business getting what it got was the San Antonio Spurs. In exchange for George Hill – an average guard still on his rookie contract – the Spurs got Kawhi Leonard, who posted a PAWS/40 of 13.02, which was good enough for 4th amongst all drafted players. The really amazing part about the trade is that they still have a strong guard rotation (Parker, Ginobili), it shores up the Spurs’ weakest position (small forward), and takes minutes away from the unproductive Richard Jefferson. Now the Spurs can trot out a starting lineup of Parker, Ginobili, Leonard, DeJuan Blair, and Tim Duncan, and try for one last ring before Duncan retires.
4. Who made the biggest reach?
Canadian Tristan Thompson had a PAWS/40 of 8.18, which put him 51st out of all the players who were drafted, and yet the Cavs selected him at #4. While I’m glad they did – it certainly removed any chance that my favorite team, the Raptors, would draft the hometown boy one pick later – it was not a good pick. They made the right choice by picking Kyrie Irving with the #1, but man…there were much better options available at #4. Honorable mentions go out to Washington (Jan Vesely, ranked 66th) and Detroit (Brandon Knight, ranked 65th).
– Devin
Wages of Wins Journal 
cheetohills
June 26, 2011
As a Bullets fan I liked who we picked up in the draft; an exciting young player and more mature productive players.
Looking for more meta analysis on this topic I ran across this post from RCS. Others might find it informative as well.
http://www.realclearsports.com/blognetwork/rcs_sidelines/2011/06/nba-draft-report-card.html
BTW, no I didn’t post it just because my Bullets scored very high there.
Devin Dignam
June 26, 2011
cheetohills:
Thanks for that link; it’s really interesting. Once my full piece is ready you’ll be able to compare my grades to those of that article…but the Wizards (yes, they are still called the Wizards for the time being) will be getting a grade in the F-C range. Some other teams whose meta-ratings are way out of whack in that article are the Pistons, the Suns, the Heat, and the Clippers.
Frederic Bush
June 26, 2011
I think that the Spurs trading Hill for picks means they’re convinced that a lockout will kill next season. He was a popular, productive player and still a bargain since as you point out he was on the last year of his rookie contract. However, after next season he will no longer be a bargain, so if there’s no next season…
Tyler Reed
June 27, 2011
Heh. Doesn’t history tell you this analysis means absolutely nothing? Miami sure did win the 2008 draft. Is this serious or just for fun?
Tyler Reed
June 27, 2011
Okay, I guess you covered that. My bad.
doclinkin
June 27, 2011
Here again we see the PAWSys hugging tight to their pet stat in staunch defiance of reality.
Great, so Chris Singleton and Tristan Thompson can’t play. Okay. So help me out here:
http://kenpom.com/index.php?s=RankAdjDE
Looks to me like FSU and Texas were the #1 and #2 most efficient defensive teams in the NCAA last year. They forced opponents to miss at a remarkable rate, offering them only bad shots or none at all. Any armchair scout who watched games carefully saw quite clearly the reason why: Singleton and Thompson were all over the court, defending their man, your man, acting as one-man zones, floating in that middle ground to deter the dribble-drive, hopping to the outside to hedge the pick and roll then collapsing to defend the interior.
Singleton most notably had no one playing behind him and essentially singlehandedly (so to speak) accounted for a majority of those errant caroms. Thompson had Jordan Hamilton behind him cleaning up the misses and padding his rebounding totals. ( I suspect JorHam shows up tall in his PAWS score).
No doubt rebounding is important. Ditto efficient scoring. But box score stats show a skewed slice of real game effect. Win Score god Joey Dorsey is not better than his Memphis teammate Derrick Rose. Never will be, never was. If you ever thought so you got to question your numbers, not your eyeballs.
doclinkin
June 27, 2011
Effectively if you are challenging opponent perimeter players and playing free safety in the midrange you will see anemic rebounding numbers since you will be out of position often to collect the misses your actions have forced. Since Win Score is weighted to emphasize rebounding (and doubly so for traditional interior boardsmen like Power Forwards in Positional Adjustments) these players will look weak.
But this supposed weakness defeated PAWS champions like Kenneth Faried (whom I adore like a favorite nephew) and allowed a team like FSU to play winning ball despite the 134th best offensive output in the Nation.
I have no good stats for Eurosquads, I’d love to see adjusted on/off numbers for Defensive Rebounding % for instance, but having watched a few games of Vesely in action (full Partizan games are available on Youtube) he too was extremely effective at sealing his man to allow others to rebound. I’m told video analysis of his play shows that opponents posted a 40% FG% against him and one in every three possessions of his match-up resulted in a defensive forced turnover. His role was as a perimeter forward, crashing the lanes on pick and roll, dribble-drives and alley oops on offense (or quick dunk putbacks) and on defense often was a perimeter wet blanket far outside paint. Regardless, given that he played a significant role with relatively heavy minutes on a deep team, and given that his team was a championship squad, I can’t see that his measured ‘win’ effect counted against the team’s ability to actually, you know: win.
I’m curious to see if these players can adjust to different roles. But my eyeballs say these are smart savvy athletic feisty players who are performing in the roles their coaches have detailed for them, and who are actually winning in the standings where the numbers truly count. I suspect they’ll be able to adjust to new roles as needed, much like say, that John Wall character (about whom you may remember, I was proven right as well. I’m Dizzy Gillespie here when it comes to tooting my own horn). I also suspect that the rebounding totals and PAWS of a player like JaVale McGee will spike upwards playing behind guys like Singleton and Vesely. More missed shots mean more rebounds to grab. Even if he’s often out of position for actually defending his match-up.
Daniel Petersen
June 28, 2011
The Hill trade was great– Gary Neal was .059 WP48 in his rookie year, which is indicative of a player who will likely be average during his career. The Spurs have players who produce at an above-average rate in Duncan, Ginobili, Parker, Blair, McDyess, Splitter and Bonner (who alternates below-average and above average years for some reason), as well as Novak and Green who are likely being brought back. Jefferson ended the year with a .070 WP48, which isn’t close to the disaster everyone says it is. 3.6 wins at $8.4M isn’t as bad as his 2.something wins at $14M last season. He’s not a bargain by any means, but he’s no longer one of the most overpaid players in the league. If Kawhi Leonard ends up being above-average and Jefferson continues to get better acquainted with the Spurs system and gets back to being over .100 again, the Spurs could have ELEVEN above-average players next season.
60 wins and another #1 seed? Yes please.
arturogalletti
June 28, 2011
Daniel,
The only problem is that they need to be top heavy to succeed in the playoffs. My take is that Pop has a similar model to mine handy and is taking a shot at Leonard being a Star/Superstar piece that he needs to go to the next level.
Tom Mandel
June 28, 2011
Arturo — once again the usual error of employing a statistical measure for sorting a population to critique *individual* decisions. Try going back to the ’09-10 draft using your measures. Would you have drafted Cole Aldrich before John Wall? How about Bryan Zoubek? How about Marqus Blakeley?
You (that is, WoW) *would* have picked Evan Turner in front of John Wall — who had the more productive season? Ditto Cousins in front of them both — but he had a worse season than both of them. Ditto Wes Johnson in front of Wall at least (I’m not taking the time to double-check these but I’m pretty sure of that). How’d he do?
How’d you like Paul George measured as an SF (wch he played in college)? Look at nerdumber for his rookie wp48. Landry Fields presented quite good SF numbers in college numbers, but nowhere near Wes Johnson, nor near his rookie NBA performance which couldn’t have been predicted.
Not that WoW analysis didn’t have some wins — Damion James was a heck of a bargain near the bottom of round 1. But where would you have picked Trevor Booker, who I believe was one of the top 10-12 WP48s among rookies?
Every tool in every field attracts people who use it for what it offers and also attracts true believers who think it is a universal solvent, a telescope into the future, or some other secret sauce. WoW is no different from any other in that regard.
Of course, WoW *is* useful in many circumstances — but it’s only a roll-up of box score stats subjected to regression analysis via stat software. The measure of its goodness is whether it’s a better roll up than EFF or PER — better in a way that can be tested. And it is that; it’s the best single to sort a population of players by productivity.
But it’s *not* something you can use to rank draft prospects. You don’t *ignore it* — one reason I liked Booker, for example, is that his WoW numbers were much better than most other availables at #23. But you don’t live and die by it!
Man, it’s tiresome to have to point this out repeatedly. Please do respond to the ’09-10 draft questions above.
(This is not to say that I disagree about Denver’s draft, btw — I *love* Faried and like Hamilton a lot)
Tom Mandel
June 28, 2011
Arturo — you write: “For the most part… players with a PAWS of 12 or higher usually end up being good NBA players.”
Uh huh. Is that an empirical claim? That is, are you actually stating that most NCAA players w/ those numbers end up as good NBA players?
Are you familiar w/ the concept of the “cemetary?” I.e. “most” of those players *aren’t drafted.* Why? Because they don’t pass the GM filter. But given that the GM filter is demonstrably close to a random walk (according to you and WoW), we must conclude that “most” of those undrafted guys would in fact turn out to be “good NBA players”.
Are you sure you want to make that claim? And if you *don’t*, what does that say about the value of the measure for this set of individual decisions?
Just how good an NBA player do you think Bryan Zoubek would be? Again, please *respond*; simply manipulating some stat software is *not* analysis. You’re a very smart guy, so you know that.
Tom Mandel
June 28, 2011
doclinkin knocks it out of the park w/ most of what he says (not a surprise). Worth mentioning however that Singleton actually has quite good rebounding numbers for a 3.
Assuming both Singleton and Hamilton are measured as 3s, CS’s deficiency in comparison w/ JH consists of *exactly one stat* — namely that he committed twice as many fouls. Remove both guys fouls, and Singleton measures slightly *better* than Hamilton.
Might Singleton’s foul rate possibly reflect his role in team defense? Does the single PAWS number reveal or obscure something in this individual comparison? Again, Arturo — please respond to the point.
Manipulating stat software is not analysis. It’s good to have the info it provides, but that’s as far as it goes.
Tom Mandel
June 28, 2011
Ooops, reading over my first post I see I have “nerdumber” for “nerdnumbers” — I *like* nerdnumbers. Please don’t take a slip of the fingertips for an intentional slight. Not so meant.
arturogalletti
June 28, 2011
Tom,
Devin is the author here but I did write multiple posts on this:
http://arturogalletti.wordpress.com/2011/06/21/early-returns-on-the-2010-nba-draft/
https://dberri.wordpress.com/2011/06/23/a-note-and-the-goods-ranking-the-2011-draft-prospects/
I’d still pick Turner over Wall. Wall has a higher WP48 (.099 vs .086 for Turner) but if I account for team quality they’re functionally the same. Wall also shows up as the worse defender in the league when I look at opponent production. This is one of those that I feel is still to be determined in the long run.
Paul George was a no. Fields would have been a yes as an SG (I screwed up there). Zoubek was a no.
arturogalletti
June 28, 2011
Tom,
Again. Devin not Arturo. I take a more holistic approach when I project and my numbers bear me out (take a look at the links the model historically does very well ). Except for freaking Michael Beasley.
arturogalletti
June 28, 2011
Um, manipulating stat software can very much be analysis. When done by capable hands of course.
Tom Mandel
June 28, 2011
Arturo — stupid me for not addressing Devin (apologies, Devin!).
Sorry, but “Zoubek was a no” doesn’t work. Either this is the measure to use or it is not the measure to use in making a single-case decision. If it’s just one thing to look at, and one takes it into account to a greater or lesser degree case by case, then there’s no point to the slam on the Wizards, is there?
If by “a more holistic approach” you mean you don’t rely on WoW analytics, then where are we — statistics plus… intuition? That’s not a “model” (tho of course it’s a method).
The “model historically does very well” means that *over a large population* it sorts better than some other model. “Does very well” can have no other meaning. But we are not discussing how it does over a large population or whether it’s better or such a population than some other model. We’re discussing whether one can say w/ *any* confidence (in the technical sense) that the Wizards’ 3 choices constitute one or another kind of draft — good, brilliant, awful, etc. That’s the nature of the judgment made by the piece.
If on the other hand, you mean something like “if the model is reliable, then the Wizards had a bad draft” I hope you realize that this statement asserts *nothing* whatever. That is, it reports in indirect language that “if I’m right, then I’m right.” To put it slightly differently, the propositional logic of an if-then statement depends *entirely* on the truth of the “if” part of the statement.
*All* attempts to use a statistical model — however complex — to predict a specific future reality-state have a similar problem; take the time to read Taleb’s book. This one is just particularly ridiculous.
Again, that’s not to ding WoW. It is what it claims to be: a measure that does a better job than any other in attributing contribution to team wins appropriately among the players on the team. It’s better than PER, it’s better than… etc.
What Devin has done in this post, what Dave did last year in mocking the pick of Wall over Cousins, what others have done repeatedly along the same vein constitute nothing more than amusements. They give pleasure to a group of true believers.
I *never* see anything approaching a *retraction* for example. Mistakes never seem to indicate any kind of problem with the model. We are expected to believe that Nick Fazekas would be an all star if only he could land a spot on an NBA team, and that Joey Dorsey already is one. What a way to misuse something which does have real, if modest, intellectual value.
Devin Dignam
June 28, 2011
Tom:
I think you’re being a bit picky here. I have included my disclaimer at the top – you did catch that, right?
The measure – any measure – will never be 100% perfect. For example, PAWS/40 doesn’t take into account age (yet), defense beyond blocks/steals/rebounds (yet), injuries, and psychological factors (such as “coachability”, ability to deal with stress, interactions with teammates, etc). And even if it did, it still wouldn’t be perfect. But PAWS/40 is the best way to quickly assess draftee productivity, so it’s the measure that I decided to use. What else would you have us examine? PPG? PER? Draft combine results?
I explicitly stated that I was using PAWS/40 was the method I was using to grade the draft, and I also explicitly stated that the correlations are not perfect. Could I mention anything else in the disclaimer that would make you happier?
Arturo’s “more holistic approach” does indeed rely on WoW analytics, and I’m sure it does even better than PAWS/40, as it uses height and age data in addition to position and Win Score (which are included in PAWS/40). Check out his awesome links. I’m sure he’ll work on some kind of draft review at some point.
I think everybody has a responsibility to use the best methods available to them to answer their questions. That is what I attempted to do with this post. I have not claimed that it is infallible, and I have mentioned that it gets things wrong sometimes. I am no “true believer”, and I really don’t think you’ll find any people who rely on NCAA PAWS/40 unquestioningly.
reservoirgod
June 29, 2011
I think Tom Mandel makes a good point that the numbers need to be presented w/ some context (at least, that’s what I took away from his comment) but I also think Devin makes a good point that Tom’s being a little picky. After all, the post is titled “Quick Takes…” so the expectation for context was set pretty low.
Devin doesn’t make the claim that the article is an in-depth analysis of the draft, and in fact he says otherwise. The article has value because it uses some objective measure to support the writer’s opinion of the draft as opposed to other articles handing out draft grades based on groupthink or what NBA scouts and the Dead Basketball Poets Society have told them.
And Tom, let’s be honest… the Wizards stink. I don’t think Ernie Grunfeld deserves the benefit of the doubt on anything.
As for Doc Linkin & the defense argument… defending players in a garbage ACC is one thing & defending players in the NBA is another. Let’s see Singleton do it in the NBA first. It’s not like there have been any studies measuring how defense translates from the NCAA to the NBA.
doclinkin
July 3, 2011
Forgot to check for follow-up.
As for Doc Linkin & the defense argument… defending players in a garbage ACC is one thing & defending players in the NBA is another. Let’s see Singleton do it in the NBA first. It’s not like there have been any studies measuring how defense translates from the NCAA to the NBA.
Sure there have. Depends what you’re measuring of course, and how you define defense, — but hell, if we had effective isolated numbers for defense we wouldn’t have any awkward clunky stats like Win Score trying to approximate and replicate on-court effect..
(FWIW–> if you’re looking at box score stats alone from NCAA to NBA: blocks per minute seems to correspond at a 93% rate, rebounds: 83%, steals just under 60%).
The real concern with defensive specialists seems to be that they tend not to earn significant playing time. Studies show the stat with the strongest relationship to minutes played seems to be points per minute. Not eFG, nor any nuanced aggregate stat of player effectiveness, nope: raw scoring. Hell I don’t need to tell the PAWSies this of course, that’s likely chapter one in the book and raison detre for the site and stat.
But right, the concern with Singleton is that his (current) sub-optimal scoring efficiency will be a drag-chute on his potential for play time. I’m not that concerned, he interviews as a hard-working determined kid and while his shot is slow, and his hands are large, he displays decent form on his jumper. I don’t doubt he can improve once he is no longer the sole focus of opposing defenses.
Still as for doubts that he can defend at the NBA level I have a tough time finding grounds in reality for that. I’m not terribly concerned that the supposedly weak ACC stunted Singleton’s growth or shaded him from exposure. In his three year tenure the ACC has consistently posted top-ranked offensively efficient squads (in Duke, UNC, Miami most notably). And Florida State’s strength of schedule (in terms of opponent offenses) has been in the top 10% of D1 schools. In that time Florida State jumped from the 41st best defense to the 12th (in Singleton’s freshman year, coming off the bench, earning more minutes late) to 1st in the nation for both his sophomore and junior years.
It is a truism spoken by NBA coaches (most notably Greg Popovich and his descendants) that in order to defend you need one strong interior presence and one multipurpose long perimeter defender. One guy to force misses and degree of difficulty shots, and one guy to deter interior attacks and to grab rebounds from missed shots.
(Tough to find statistics for this but here’s one study showing the relative adjusted +/- by position.
http://basketball-statistics.com/atwhichpositionsaregreatplayersthemostimportant.html
Small Forwards show up with the most pronounced overall effect, followed by post players (PF/C’s)).
The most successful teams often follow this model. I suspect the Wizards will begin to see dividends from the pick-up of a player like Singleton who can play that perimeter defensive role, even if the team hasn’t yet developed that interior brute who can captain the defense and be relied on as a rocksteady presence more than a highlight waiting to happen at either end of the court (JaVale McGee, I’m lookin at you kid).