When somebody says David Harrison was a bad draft pick, is that accurate? At what point in the first round does is stop making sense to expect a starter? What is the value of tanking? Perhaps more importantly, what is the reliability of tanking? The following post may or may not answer these questions, but with the Draft Lottery taking place tonight, we’re going to take some time to try to understand what kind of return teams have historically gotten from draft picks.
For this analysis, I’ve collected data on every draft class from 1977 through the current one, encompassing all regular season activity from the 1978 season to the recently finished 2010 season. During that time frame, over 3,450 draft picks made, and over 17,00 of them played in at least one NBA regular season game.
To make it more relevant to today’s two-round system, I’m going to limit the study to the Top 60 players taken in each draft. For some, that will mean dipping into the third round, but for all of the classes between 1988 and 2005, it will mean taking the entire two rounds, though those drafts had fewer than 60 selections. This drops the sample to 1,992 draftees, of whom 1,553 played in at least one NBA game.
Today, we’re going to use the data available from the inimitable Basketball-Reference.com to try to crank through some numbers. Later, in Part II , we will go through all of the awards and honors bestowed on the players in this sample.
Number Crunching
In order to do a statistical analysis on this many players, I needed something that approached a unifying number or metric. For this analysis, I dusted off an old metric developed by Martin Manley in the late 1980s called “Production Rating.” It is calculated as follows:
Production Rating (PR) = (Points + Rebounds + Assists + Blocks + Steals – Turnovers – Missed Field Goals – Missed Free Throws)/Games Played
To update this metric a little I’ve made two adjustments to it:
- Pace – I have basically adjusted all of the PR’s to a per 100 basis. As a shortcut, I used the Pace Factor for the player’s team for this adjustment. For example, Danny Granger’s 2010 numbers were “played at” 97.1, so they were multiplied by (100/97.1). It’s not perfect, but it’s sufficient for this purpose.
- Reliability – Essentially, this is just a way to adjust for games missed. The net effect is to treat the games missed as a zero (0) PR. For career reliability, I put a minimum number of years at five (5) years. This clearly doesn’t impact players whose draft classes haven’t been in the league long enough, but it is meant to penalize players who played shorter than average careers. An example would former Pacer Kenny Williams, who only played for four years. His 260 games would be divided by 410, instead of 328. If a player played five or more years, he was not penalized for “missed” years. For example, neither Michael Jordan nor David Robinson were penalized for the full seasons that they missed either at the beginning and middle of their careers, respectively.
There are flaws in this system. It will overrate stat stuffers like Shawn Marion or Troy Murphy. It will underrate players like Scottie Pippen, Joe Dumars and Shane Battier, but, hey, so does PER. Overall, however, I don’t believe that this analysis is telling you (or me) any lies.
If you want to see how others have done this, Tom Haberstroh used EWA in the D.R.A.F.T. Initiative on ESPN, and Roland Beech of 82games.com used a very simple rating combining Points, Rebounds, and Assists per game.
Breaking Down the Draft
To jump start the numbers discussion, let’s start with this chart:
This shows the simple average for the Adjusted Production Rating per 100 (AdjPR100) for each of the top 60 draft slots over the last 33 years. I’ll explain further what the AdjPR100 number “means” in a little bit, but this gives a good visual representation of the the way the draft flows in terms of pick value (using production as a proxy for value).
The #1 pick stands almost head and shoulders above all others, with its 16.3 average a full 20% higher than the second most productive slot (#3). The rest of the top five are grouped relatively closely together before decent drop off to the second half of the top 10. The second half of the top ten is interesting in that the #10 pick actually has posted a slightly higher average (9.4) than the #6 has (9.3). Flowing to the right, you see the continued decline, but with a flattening curve as we approach #60. As we’ll see later, the averages in the latter half of this sample will be greatly diluted not just by below average performers, but by players who have never seen a minute of NBA play.
From a broad perspective, the visual could be said to tell us what we already know: the earlier the draft pick, the better. However, my eyes shows me subtle fluctuations that call for further investigation. In order to avoid dragging you through each of the 60 picks, I will put the picks into groups. For the “First Round” picks, or Top 30, each group will contain 3 picks (1 to 3, 4 to 6, etc.). For the “Second Round,” each group will contain 10 picks. This smooths the curve a little bit, but I will come back and discuss specific draft picks, especially whichever slot the Pacers end up having.
Seeing Stars
Within each of these Draft Groups, I going to stratify the AdjPR100 using a 5-star rating system. Here is a brief explanation of the ratings:
5 Stars (21 and above) – The very best, most productive players. Players who attain this level over a full career become All-Stars, All-NBA players, or even MVPs.
4 Stars (16 to 21) – Almost all of these become either All-Stars or All-NBA players. There have been players in this group earning MVPs (Allen Iverson, Steve Nash) and others elected to the Hall of Fame.
3 Stars (12 to 16) – Generally speaking, these would be players who represent good starting material, but there have been plenty of All-Stars and All-NBA nods in this group. They might not be perennial, but they can peak at that level. There is one Hall of Famer at this level: Joe Dumars, whose defensive contributions are underrated by this (and most) rating systems.
2 Stars (8 to 12) – Good, solid players. Starters in some situations, key reserves in others. There are some All-Star and All-NBA appearances, but not regular ones. Drazen Petrovic was a Group 2 player who made the Hall, but that was based on his European career, as well. You’ll also see some players here like Derrick McKey, Paul Pressey or Michael Cooper, whose defense would be strong enough to put them ahead of some higher rated players in the real world.
1 Star (3 to 8) – Some rotation players, some fringe players. You have one All-Star (Jayson Williams), and a handful of guys who made All-Defensive teams as specialists — Thabo Sefolosha, Bill Hanzlik, T.R. Dunn, to name a few.
0 Stars (0 to 3) – Anywhere from fringe players to players who haven’t gotten their shot to outright busts. This group also includes the 369 draftees who have not played a single game in the NBA. Among those are Blake Griffin and Ricky Rubio, who are expected to earn much higher ratings once they actually suit up and step on the floor.
As noted in the explanations, this is not perfect, but I believe it’s functional. While you could certainly argue (and I probably would, as well) that Scottie Pippen should be a 5 Star, instead of a 4, or that Joe Dumars should likewise be a 4 Star, but the rating system as a whole, in my opinion, makes sense. Here’s a look at the distribution across the 1,922 draftees in the study.
This shows a very skewed distribution towards the bad. I believe this is accurate, because the majority of the players who wander through the league are fringe players. In any given season, 450 players see the floor for an NBA team. There are only 150 starting slots, and the vast majority of coaches only give meaningful minutes on a night basis to eight guys. About 5% of the players in a season will become All-Stars, and about 3% will be named to one of the All NBA Squads. With this perspective of the overall force structure of the NBA in mind, it’s time to start marching through the draft groups.
Picks #1 to #3
Number of Draftees: 99
Top-Rated Player: LeBron James (#1, 2003) 29.57
Lowest-Rated Player: Chris Washburn (#3, 1986) 0.52
Never/Has Not Played: 2 (Len Bias, picked #2 in 1986; Blake Griffin, picked #1 in 2009)
For a complete list of all 99 players, click here
Star Distribution:
Without question, this is the place to be in the draft. Unless you just completely screw up (I’m lookin’ at you, Joe Dumars), you’re going to get a starter and more likely than not, an All-Star. (At this point, I should make the caveat that all drafts are different, simply because they have different draft classes. Though I may lapse into shorthand about chances and expectations, I am not implying that every #1 pick offers the same opportunity. History is never a perfect predictor of the future, but it can give you an idea of when to be happy and when to be disappointed.)
You’ve got 11 of the 24 Five-Star guys in this group, including guys like LeBron, Michael Jordan, Timmy Duncan and Magic Johnson. The vast majority of these guys find ways to be productive members of the NBA society. Whether those guys are living up to the hopes and dreams placed on picks made so high is another question, but this is where most of the best players are taken.
There are also some real tragedies here. Jay Williams played moderately well as the #2 pick for Chicago, before his career was ended by a motorcycle accident. I can still remember the day they found Len Bias’ body. The league and the city of Portland are fretting over whether Greg Oden will ever be healthy enough to fulfill his promise. Blake Griffin became yet another chapter in the star-crossed history of the Clippers by losing his rookie year to injury.
In any case, this is where MVPs and All-NBA players are “born” into the NBA, so you better not drop the ball.
Picks #4 to #6
Number of Draftees: 99
Top-Rated Player: Kevin Garnett (#5, 1995) – 27.18
Lowest-Rated Player: Russell Cross (#6, 1983) – 0.40
Never/Has Not Played: 1 (Ricky Rubio, #5 in 2009)
For a complete list of all 99 players, click here
Star Distribution:
Still plenty of good “gets” here, but the distribution slides ever so slightly to the right. The 63% hit rate of 3-Star or better players in the Top 3 picks drops to only 45% here, meaning that we’ve just that quickly moved into the area where their have been more 2-Star or less players than 3-Star or better.
Besides Kevin Garnett and Larry Bird, other greats include Scottie Pippen, Charles Barkley, Chris Paul, Dwyane Wade and Chris Bosh. The 2010 Rookie of the Year, Tyreke Evans, was taken by Sacramento with the fourth pick. You’ve got a decent shot at something special here, but you can’t quite count on it. Drafting here should get you a solid starter, with a downside of a rotation player.
What failure looks like here is William Bedford, Nikoloz Tskitishvili, and, yes, Pacer fans, Jonathan Bender. It remains to be seen whether it will also look like Minnesota’s head scratching decision to take two point guards (Ricky Rubio and Jonny Flynn) with the 5th and 6th picks last year. Flynn put up a respectable 2-Star rating (11.3) this past season, but Rubio spent the season in Spain, and Minny slogged to 15 wins. On the bright side, they’ve got a 56% shot at at Top 3 pick this season and a 100% chance at a Top 5 pick.
Picks #7 to #9
Number of Draftees: 99
Top-Rated Player: Dirk Nowitzki (#9, 1998) – 25.73
Lowest-Rated Player: Patrick O’Bryant (#9, 2006) – 0.82
Never/Has Not Played: None
For a complete list of all 99 players, click here
Star Distribution:
Before I start discussing the overall distribution of this group, I wanted to point out something I just noticed. In each section, I’ve listed the highest- and lowest-rated player (who played) for the group. The players in each group listed as lowest — Chris Washburn, Russell Cross and Patrick O’Bryant — all have one thing in common. They were each drafted by the Golden State Warriors. You can take that for what it’s worth.
Moving on to the ratings, this group has only produced two 5-Star guys — Nowitzki and Shawn Marion — and Marion should probably be noted as an outlier — an overvalued player who is product of the rating system. There are almost as many 4-Star guys (11) as in the #4 to #6 group (12), and they include Jack Sikma, Andre Iguodala, Amare Stoudemire and Andre Miller. The 3-Star, 4-Star and 5-Star players make up a touch over one-third (35%) of the total, while there are only a couple more busts.
The busts here aren’t particularly memorable. Besides O’Bryant, there are guys like Rafael Araujo, Bo Kimble and Ed O’Bannon. Perhaps the most notable league-wide would be Joe Alexander, who suffered the ignominy of being the highest drafted player not to have his third-year option picked up. Pacer fans will think of George McCloud and Ike Diogu, both failed experiments in Indy. McCloud, however was able to salvage some semblance of a career as a fringe rotation player, falling just short of making the 2-Star rating.
Picks #10 to #12
Number of Draftees: 99
Top-Rated Player: Brook Lopez (#10, 2008) – 21.47
Lowest-Rated Player: Yaroslav Korolev (#12, 2005) – 0.07
Never/Has Not Played: 1 (Fran Vasquez, #11 in 2005)
For a complete list of all 99 players, click here
Star Distribution:
The #10 to #12 grouping actually straddles one of the step downs in the draft. If you refer back to the very first chart in this piece, you’ll see drop-offs from the #1 pick to the rest of the Top 5, from the Top 5 to picks #6 through #10, and then into the late lottery. As a whole, the picks #10, #11 and #12 have been 50/50 between producing solid or better players and producing fringe players and busts. However, that’s not something that gets spread like peanut butter across these three slots.
There is a marked difference between pick #10 and the #11 and #12 slots. In fact, the draft history for the 10 pick has been as good or better from both an Average AdjPR100 basis and a Star Distribution than picks #6 through #9. Looking down the list, you see some very nice players indeed — Brook Lopez, Paul Pierce, Joe Johnson, Jason Terry, Caron Butler, Brandon Jennings, Eddie Jones, Jeff Malone and Andrew Bynum.
Where the #10 pick has ten 3-Star, 4-Star or 5 Star players out of 33 picks, the #11 and #12 picks managed only 12 out of their 66 chances. Reggie Miller appears (at least through my Blue-and-Gold-colored glasses) to clearly be the best player among the latter two picks, standing as the lone 4-Star guy in the group. Pearl Jam might make an argument for Mookie Blaylock, but I’d win. Others of note are Fat Lever, Kevin Willis and Cedric “Cornbread” Maxwell. Jason Thompson is the highest-rated youngster in the group.
The flops here are hardly worth mentioning. As I peruse the names nobody jumps out as someone who would have had high expectations, but failed. Perhaps Jerryd Bayless, but he just really hasn’t gotten a chance yet and has shown flashes that he might still be a very good player in this league. Mostly, I think things like, “Trajan Langdon? Really?” I do, however, enjoy looking at the 2005 draft, where Orlando went for Fran Vasquez and the Clippers drafted Yaroslav Korolev, thus aiding and abetting Larry Bird in being able to end trade talks with another team that evening by saying, “I gotta go. I’m going to draft Danny Granger.”
Picks #13 to #15
Number of Draftees: 99
Top Rated Player: Karl Malone (#13, 1985) – 27.83
Lowest Rated Player: Scott Haskin (#14, 1993) – 0.24
Never/Has Not Played: (Frederic Weis, #15 in 1999)
For a complete list of all 99 players, click here
Star Distribution:
Karl Malone! Kobe Bryant! Clyde Drexler! Steve Nash! TROY MURPHY!!!!!! Holy Counter-Intuitive, Batman! Teams shouldn’t tank for the Top Three, they should be targeting late lottery. That’s five, count ’em, five MVP awards and four (including future) Hall of Famers right there. And the cherry on top is T-Murda, baby!!!! Get down, get down — uhhh — get down, get down.
Perhaps it’s because it has the 13th pick, but this is just a weird mix. You’ve got four bona fide Hall of Famers tucked in with some quality players (Tim Hardaway, Al Jefferson, Dale Davis and Thunder Dan Majerle) then poured in a vat with a whole buncha blah. This is the point in the draft where it really no longer makes any sense at all to expect to be able to grab a starter. In fact, history shows an almost 2-to-1 edge for low rotation, fringe and bust players over significant contributors.
In fact, we’re going to see the draft kind of flatten out over the next 10 picks or so. Of course, every draft is different, but most aren’t as different as we like to pretend. To me, this is an area where GMs should begin to feel a little more comfortable gambling. They should feel more comfortable trading back, and this is an area where they should make “their” mistake. Now, I don’t mean that they should screw up. I simply mean that since there’s more inherent risk in these picks, that they should go with the guy they like, even if he’d be considered a reach.
That type of reasoning could be used to explain Bird’s selection of Hansbrough last summer. I was not a fan at the time, but even in the little time that he played last year, I could see some of what Bird and O’Brien would see in him. (And, no, it’s not that he’s white.) By the same token, this reasoning could also be used to argue that every GM who passed on DeJuan Blair was a pinhead. Yes, the knees were risky and, yes, there were and — will be — times that his size will hurt him in the league, but the talent is there. Perhaps #13 to #15 in last year’s draft was too high, perhaps not. Still, he should have gone somewhere in the first round.
Picks #16 to #18
Number of Draftees: 99
Top Rated Player: John Stockton (#16, 1984) – 21.74
Lowest Rated Player: Luther Wright (#18, 1993) – 0.03
Never/Has Not Played: 2 (Rod Griffin, #17 in 1978; Troy Bell, #16 in 2003)
For a complete list of all 99 players, click here.
Star Distribution:
Some teams have definitely struck gold in this part of the draft, but that’s something of a misleading vivid. John Stockton and Joe Dumars are the last of the enshrined Hall of Famers from the first round of these draft classes. After that, you’ve got a pretty nice collection of names. Statistically, Josh Smith heads up the list and is joined by Shawn Kemp, David West, Hedo Turkoglu and James Posey.
The Pacers have gotten a lot of mileage from this part of the draft (though not all were selected by the Pacers). Danny Granger is the most celebrated, but other Pacer connections in this area include Jermaine O’Neal, Mark Jackson, Vern Fleming and Roy Hibbert.
However, the big thing to say here is that this is the first group where the zero (0) Star rating has the most players. Around the late teens is where I start to understand teams selling draft picks, and start to scratch my head at the willingness of fans and GMs to trade established rotation players and even starters for these picks.
Picks #19 to #21
Number of Draftees: 99
Top Rated Player: Larry Nance (#20, 1981) – 19.31
Lowest Rated Player: Monti Davis (#21, 1980) – 0.01
Never/Has Not Played: 1 (Larry Knight, #20 in 1979)
For a complete list of all 99 players, click here.
Star Distribution:
More of the same here, though there’s a lot of reason for Boston fans to be excited about Rajon Rondo, taken 21st in 2006. Lots of role players here, and though the zero (0) Star column is no longer the biggest, that’s primarily because the 1-Stars are stealing from the higher-rated tiers.
A couple of random Pacer-related thoughts here. First, considering Shawne Williams over Rondo, Hansbrough over guys like Ty Lawson, Eric Maynor and Darren Collison, and the overwhelming apathy that I (and most of the rest of the league) feel towards Brandon Rush, perhaps I’m not so comfortable with the idea of Larry making “his” mistake. Second, for all of the clamoring for a Dale Davis to put next to Roy Hibbert, it strikes me that a Larry Nance-type (or at least the Larry Nance I remember) would probably be a much better fit.
Also, Jeff Foster, drafted 21st, is in this group as one of the 2-Stars guys. He strikes me as a good illustration of what a good 2-Star player is: reliable, but limited role player capable of making a contribution to a good team.
Picks #22 to #24
Number of Draftees: 99
Top Rated Player: Reggie Lewis (#22, 1987) – 15.93
Lowest Rated Player: Tom Sewell (#22, 1984) – 0.00
Never/Has Not Played: 5
For a complete list of all 99 players, click here.
Star Distribution:
Arguably, this has a better distribution than the last group, with 33% being 2-Star guys or better, as opposed to 23% for picks #19 to #21. Another tragic story here with Reggie Lewis, whose death cut short a promising career (and life, which would be infinitely more important to his loved ones). Though only a 3-Star rating for his career, his last two season of AdjPR100 were over 20. Drafted in the Year of the Reggies (the 1987 draft class included Lewis, Miller and Williams), many thought he was becoming (or already was) the best of the three. This, of course, is nothing more than blasphemy.
Some other fun names here. Norm Nixon, who I remember for three things: playing for the Lakers, blowing out his knee in a softball game and being married to Debbie Allen. Tayshaun Prince, who the Pacers actually drafted instead of Freddie Jones, and most certainly did not block Reggie’s lay-up the year the Pacers won the NBA Title. Arvydas Sabonis, who I believe made David Robinson cry in the 1988 Olympics. Some good role players, tempered by the almost 40% bust rate.
Picks #25 to #27
Number of Draftees: 99
Top Rated Player: Vlade Divac (#26, 1989) – 16.68
Lowest Rated Player: Ron Moore (#25, 1987) – 0.03
Never/Has Not Played: 6, including Vern Fleming’s twin brother Victor
For a complete list of all 99 players, click here.
Star Distribution:
Besides Vlade and his beard, you also find Dennis Rodman here. The Worm was a complete tool, but he was a great basketball player. He was eligible for the Hall of Fame either this year or last, and didn’t even make it as finalist. On the whole, that’s probably stupid, particularly given my opinion that the whole “first ballot” thing is an artificial construct created by some pinhead anonymous sportswriters and Hall voters to assert their control over their betters. Still, if that is how it’s going to be, I’d have to say that any guy traded — in his prime — straight up for Will Perdue probably shouldn’t expect to be “first ballot.”
The draftees in this group from the 2009 class made a nice splash, with Rodrigue Beaubois and DeMarre Carroll getting some good burn, and Taj Gibson actually making First Team All-Rookie.
Picks #28 to #30
Number of Draftees: 99
Top Rated Player: David Lee (#30, 2005) – 18.45
Lowest Rated Player: Rickie Winslow (#28, 1987) – 0.00
Never/Has Not Played: 15, though Tiago Splitter might make it 14 some day
For a complete list of all 99 players, click here.
Star Distribution:
I’m calling this the end of the first round, though some of these players were technically second round picks. Tony Parker, P.J. Brown, Gilbert Arenas, Josh Howard, Toni Kukoc and Anderson Varejao were all productive (to very good) players to come out of this portion of the draft.
Which brings us full circle to the opening question: is it accurate to say David Harrison was a bad pick? Yup. Look, it wasn’t egregiously bad, and it certainly didn’t have a huge opportunity cost. He’s lumped in with that 60% stack for the zero (0) Stars. Still, Sideshow Andy was taken with the pick immediately following our selection. At the end of the day, David sucked … and, apparently, inhaled, too.
The Second Round (Picks #31 to #60)
The big picture on the second round that most of these are nothing more than names on a sheet of paper or training camp fodder. Of the 932 “Second Round” draftees in this study, 337 of them have never played an NBA game. Over 700 of them (77%) currently stand as Zero-Star players. Only 18 players have or have had careers as 3-Star players or better. Guys like Manu Ginobili, Carlos Boozer and Paul Millsap stand out, but they are most certainly exceptions. To give you an idea, A.J. Price is currently the fourth rated #52 pick in the last 33 years, hoping to catch and pass Donald Royal, Fred Hoiberg and Rasual Butler. Of the 272 total picks between #51 to #60 in the past 33 years, AJ currently has the 24th best career.
The Pacers (probably) have two picks this year: #40 and #57. They’ll make the roster, because rookie Second Rounders are the absolute cheapest player you can get … but I wouldn’t expect much from them.
Picks #31 to #40
Number of Draftees: 330
Top Rated Player: Carlos Boozer (#34, 2002) – 18.59
Lowest Rated Player: Casey Shaw (#37, 1998) – 0.01
Never/Has Not Played: 63
For a complete list of all 330 players, click here.
Star Distribution:
Picks #41 to #50
Number of Draftees: 330
Top Rated Player: Jeff Hornacek (#46, 1986) – 16.76
Lowest Rated Player: Mille Ilic (#43, 2005) – 0.01
Never/Has Not Played: 121
For a complete list of all 330 players, click here.
Star Distribution:
Picks #51 to #60
Number of Draftees: 272
Top Rated Player: Anthony Mason (#53, 1988) – 15.26
Lowest Rated Player: Ernest Brown (#52, 2000) – 0.01
Never/Has Not Played: 154
For a complete list of all 272 players, click here.
Star Distribution:
Final Thoughts
I don’t know if this information will change anyone’s basic understanding of the draft. What this exercise has done for me is to provide perspective and specific detail to help the conversation along. This isn’t really meant to be a predictor, even though I might occasionally be tempted to speak as if it is.
If the Pacers draft at #10, history says that getting a solid starter or a rotational player should be considered a success. However, that doesn’t mean that should be the target. The reason that some good players fall to later draft picks isn’t because some drafts are deeper than others (though that can and does have an impact). It’s because talent evaluation is subjective, and not every GM comes to the table with the same view of the prospects.
Because of this, the intuitive belief that the higher the pick, the better, isn’t always true. The human factor plays heavily in how a draft flows. There will be good players available at the tenth pick. It is no stretch at all to say that at least one of them will end up being better than some of those taken before him. If a team is prepared, can understand what the players offer and how that fits into their plans, then they can change their team from the late lottery.
Still, each draft and draft class is different. We can talk about what has happened, but that may not translate directly to what will happen. One of the things about that draft, and talent evaluation in general, is that it’s far more art than science. Because of that, draft picks can’t necessarily be judged on a binary pass/fail system.
I realize that this is way too much information, but if this works right, it will set the foundation for a conversation we’ll hold in the next few weeks approaching the draft. In the second part of this series, we’ll take a look at the more subjective side – end of season awards and accolades — and hopefully that will provide even more shared knowledge for our little chat.