Exemptions Introduction NBA Not Top 10 Busts

NBA Draft Categories (Busts to Superstars)

Synopsis:  Approximately three weeks ago, Dumb and Dumber celebrated its 20th anniversary. In honor of that classic movie, I have written this post in order to simplify (or dumb down) my statistical analysis of win shares. Specifically, I have created NBA Draft categories characterized by representative players. At the same time, I have simplified the takeaways. Just to be sure, here they are. 

  1. Draft order generally is a good predictor of future success in the NBA. As such, the higher the pick, the better the player should be to avoid being labeled a bust.
  2. Starting with the 11th overall pick, the probability of being a flame-out exceeds the probability of becoming an All-Star by a margin of 2:1 (40% to 20%). For that reason, players taken outside of the first ten overall picks have been excluded as potential Top 10 Busts. 
  3.  NBA legends are rare, but not as rare as you might think.

NBA DRAFT CATEGORIES (FROM BUSTS TO SUPERSTARS)

Before beginning any research for this countdown, I thought that most potential Top 10 NBA Draft Busts would be high overall picks. It seemed logical that the cutoff point would be somewhere around 14 (i.e. a lottery pick) or 15 (i.e. the top half of the draft). After analyzing the win shares for every 1st round draft pick since 1970, I quickly realized that the falloff in production was significant even before getting to a 14th or 15th pick. In particular, there appears to be a dramatic difference between a top 10 pick and a pick taken 11th or later. Furthermore, the top 10 picks can be divided into smaller subsets with unique statistical distributions. Typically, #1 overall picks are significantly more productive than picks 2-5, who are significantly more productive than picks 6-10.

Based on this research, I was able to devise a productivity threshold below which a draft pick would have to fall before becoming a potential Top 10 Bust. Those approximate thresholds are: 16 win shares for a #1 overall pick; 9 win shares for a 2nd-5th overall pick; and 4 win shares for a 6th-10th overall pick (do you notice a trend in the numbers – 16 . . 9 . . 4 . . ?). So what the heck does it mean? To make it easier, I have taken the statistical analysis summarized in the prior post and tried to identify players with the those numbers. Specifically, I have created tables showing players whose career productivity puts them at certain milestones for each 1st round subset. In addition, I included the names of superstar players and a statistical estimate of the odds of a comparable player being taken with a comparable pick.

For instance, Andrea Bargnani is listed as a “Borderline Bust” for #1 overall draft picks based on his 18 career win shares. As such, any #1 overall pick with fewer win shares is a candidate for being declared a Top 10 Bust. Dennis Hopson (who was listed as my #8 Worst NBA Draft Pick) is also a “Borderline Bust” based on his 7 career win shares as a 3rd overall pick. Therefore, any player drafted with a 2nd-5th pick who was more productive than Hopson is safe, but any player less productive is not. Technically, Bargnani is slightly above the “border” and Hopson is slightly below, but hopefully you get the point.

With the goal of trying to make the numbers easier to understand, I have compiled the following tables with representative players used to approximate the bottom 10%, bottom 25%, median, and top 25%. Also, I have included familiar stars or superstars to give additional reference points.


As the following table shows, the players drafted #1 overall who represent the borderline picks for the bottom 10%, bottom 25%, median, and top 25% are: Andre Bargnani; Joe Barry Carroll; Larry Johnson; and Bob Lanier, respectively.

  • As someone who has averaged 15 points per game (ppg) and 5 rebounds per game (rpg) during his 8-year career, Bargnani is a borderline bust for a #1 overall pick.
  • Despite averaging 18 ppg and 8 rpg during his 11-year career, Carroll is barely above the bottom quartile (i.e. bottom 25%). Carroll is sometimes considered an all-time bust because he was involved in a trade of 1980 draft picks that didn’t go so well for his team. In particular, the Warriors traded Robert Parish and their third overall pick to move up two spots for the Celtics’ first overall pick. Whereas the Warriors used the 1st pick to select Carroll, the Celtics used the 3rd pick to take Kevin McHale. Despite this lopsided transaction (effectively a trade of Carroll for both Parish and McHale), Carroll was too productive during his career (over 12,000 points and 5,000 rebounds) to be declared a bust.
  • Larry “Grandmama” Johnson was a dominating 19 & 9 player who made 2 All-Star teams in his first five years in the league. After getting injured (which coincided with him joining the Knicks), Johnson became a 12 ppg / 5 rpg player so his overall career was only “average.”
  • Lanier, who is in the Hall of Fame, is at the bottom of the top quartile, which implies that teams with the first overall pick should expect to get a Hall-of-Fame caliber player 25% of the time.
NBA DRAFT CATEGORIES: SUMMARY OF #1 OVERALL DRAFT PICKS

Identifier

Player Win Shares Percentile Odds of Selecting a Better Player*
Borderline Bust Andrea Bargnani 18 11.6%

88.4%

Below Average

Joe Barry Carroll 36 26.2% 73.8%
Average Larry “Grandmama” Johnson 70 51.4%

48.6%

Quality Pick

Bob Lanier 117 75.5% 24.5%
Superstar Magic Johnson 156 87.1%

12.9%

Superstar

Hakeem Olajuwon 163 88.5% 11.5%
Superstar Bill Russell 164 88.7%

11.3%

Superstar

LeBron James 173 90.4% 9.6%
Superstar David Robinson 178 91.2%

8.8%

Superstar

Oscar Robertson 189 92.8% 7.2%
Superstar Tim Duncan 195 93.5%

6.5%

Superstar

Wilt Chamberlain 247 97.6% 2.4%**
Superstar Kareem Abdul-Jabbar 273 98.6%

1.4%**

* Odds that the 1st overall pick from any draft will be better than the identified player.

** Based on these percentages, the odds of selecting a similarly productive player are 1:40 for Wilt Chamberlian and 1:70 for Kareem Abdul-Jabbar.


As the following table shows, the players drafted between 2nd and 5th overall who represent the borderline picks for the bottom 10%, bottom 25%, median, and top 25% are: Dennis Hopson; Eddy Curry; Wayman Tisdale; and Byron Scott, respectively.

  • Having averaged only 11 ppg, 3 rpg, and 2 assists per game (apg) during his five-year career, Hopson qualifies as a potential bust. While his limited production put him in the running for a Top 10 Bust, there are too many other players with worse totals than those.
  • Curry was a borderline bottom quartile player based on averages of 13 ppg and 5 rpg during his seven-year career
  • With 15 ppg and 6 rpg during his 12-year career, Tisdale was an “average player” for a 2-5 pick.
  • With approximately 15,000 points, 3,000 rebounds, 3,000 assists, and 1,000 steals, Byron Scott’s career productivity was just below the top quartile for the draft subset

SUMMARY OF DRAFT PICKS SELECTED 2ND – 5TOVERALL

Identifier

Player Win Shares Percentile

Odds of Selecting a Better Player*

Borderline Bust

Dennis Hopson 7 7.3% 92.7%
Below Average Eddy Curry 22 25.9%

74.1%

Average

Wayman Tisdale 46 51.7% 48.3%
Above Avg. Byron Scott 75 72.9%

27.1%

Superstar

Bob Pettitt 136 93.1% 25.0%
Superstar Jason Kidd 139 93.5%

23.4%

Superstar

Ray Allen 145 94.4% 20.6%
Superstar Gary Payton 146 94.5%

20.1%

Superstar

Dan Issel 158 95.9% 15.4%
Superstar Jerry West 163 96.4%

13.7%

Superstar

Charles Barkley 177 97.4% 9.9%
Superstar Kevin Garnett 189 98.1%

7.4%

Superstar

Michael Jordan 214 99.0%**

4.0%

* Odds that at least one player taken with a 2nd-5th overall pick from any draft will be better than the identified player

** While someone may reasonably argue that the odds of selecting someone like Jordan with a 2nd-5th overall pick should be lower than 1:100, don’t forget that Jordan retired twice for a total of almost 5 years. It’s not unreasonable to assume that he could have added 40-50 Win Shares, bringing the odds down to 1:300.


As the following table shows, the players drafted 6th-10th overall who represent the borderline picks for the bottom 10%, bottom 25%, median, and top 25% are: Rumeal Robinson; Ron Mercer; Calbert Cheaney; and Vin Baker, respectively.

  • Robinson is probably best remembered for sinking two free throws in the final seconds of overtime to help Michigan win the 1989 NCAA Championship. Even though Robinson had a respectable 13 ppg & 5 apg season during his one year as a starter, his professional career was a bust for a #10 pick given his 7.5 ppg and 3.5 apg averages over a six-year NBA career. He may have been an NBA bust, but he already lived the dream.
Rumeal-Robinson
With his team down 69-68 to Seton Hall and 3 seconds left in overtime, Robinson had two free throws to win the NCAA Championship. He sunk both and Michigan won 70-69.
  • Ron Mercer began his career strong with averages of 17 ppg and 4 rpg for five years as starter, but his career ended disappointingly as his numbers fell to 5 ppg and 2 apg coming off bench for his final three years.
  • Calbert Cheaney is considered an average player for this subset based on his production of 10 ppg & 3 rpg during a 13-year career.
  • Before succumbing to alcoholism, Vin Baker was a four-time All-Star, two-time All-NBA, and one-time Olympian who was a legitimate 20 ppg & 10 rpg threat. Even though he was able to average 15 ppg and 7 rpg over 13 different seasons, Baker’s career was basically done after nine. Oh, what could have been!
SUMMARY OF DRAFT PICKS SELECTED 6TH – 10TH OVERALL

Identifier

Player Win Shares Percentile

Odds of Selecting a Better Player*

Borderline Bust

Rumeal Robinson 5 13.3% 86.7%
Disappointment Ron Mercer 9 22.7%

77.3%

Average

Calbert Cheaney 23 48.2% 51.8%
Above Average Vin Baker 47 73.9%

26.1%

Superstar

John Havlicek 132 97.7% 11.0%
Superstar Adrian Dantley 134 97.8%

10.4%

Superstar

Paul Pierce 144 98.4% 7.9%
Superstar Larry Bird 146 98.5%

7.5%

Superstar

Robert Parish 147 98.5% 7.3%
Superstar Dirk Nowitzki 186 99.5%

2.5%**

* Odds that at least one player taken with an 6th-10th overall pick from any draft will be better than the identified player

** The odds of selecting a player as productive as Nowitzki with a 6th-10th pick are 1:40. While picks 6-10 consistently have proven to be more productive than picks 11-14, they don’t have the same upside. With respect to these two subsets of draft picks, the distributions are different in the extremes such that Nowitzki is more rare as a #9 pick than Karl Malone is as a #13 pick even though Nowitzki has fewer win shares. Regardless, Nowitzki’s career shouldn’t be discounted as he approaches impressive milestones like 28,000 points (7th all-time) and 10,000 rebounds.


The following table shows that the players selected 11th-14th overall who represent the borderline picks for the bottom 10%, bottom 25%, median, and top 25% are:  Acie Law; Tariq Abdul-Wahad; Vitaly Potapenko; and John Salley, respectively.

  • Law finished his four-year NBA career with 700 points (4 ppg) and 300 assists (2 apg).  Ouch, that must hurt for the winner of the 2007 Bob Cousy Award as best point guard in college. Do I smell a potential Top 10 Bust brewing? Oh I forgot, he was drafted with an 11th overall pick (i.e. outside the top 10).
  • As the borderline bottom quartile pick, Abdul-Wahad was a disappointment by averaging only 8 ppg and 3 rpg averages over his six-year career.
  • Mostly coming off the bench, Potapenko was the typical 11th-14th overall pick after averaging 7.5 ppg and 4.5 rpg throughout his 11-year career.
  • Even though Salley contributed offensively with 11 ppg, he became a top quartile player for his defensive contributions, which included 7 rebounds, 1 steal, and 2 blocks per game during his ten-year career.

SUMMARY OF DRAFT PICKS SELECTED 11TH – 14TH OVERALL

Identifier

Player Win Shares Percentile

Odds of Selecting a Better Player*

Potential Bust

Acie Law 1 10.0% > 99.0%
Disappointment Tariq Abdul-Wahad 4 25.0%

> 99.0%

Average Pick

Vitaly Potapenko 15 50.0% 96.8%
Good Pick John Salley 41 75.7%

75.2%

Superstar

Clyde Drexler 135 96.2% 17.8%
Superstar Kobe Bryant 173 97.9%

10.0%

Superstar

Reggie Miller 174 98.0% 9.8%
Superstar Julius Erving 181 98.2%

8.8%

Superstar

Karl Malone 234 99.2%

4.1%**

* Odds that at least one player taken with an 11th-14th overall pick from any draft will be better than the identified player

** The odds that one team with an 11th-14th pick in any draft selects a player as productive as Karl Malone are around 1:25.  As mentioned previously, these odds might seem high relative to the 1:40 odds for selecting a similar player to Nowitzki, but the distributions of the two subsets are different. In particular, teams seem to use picks 6-10 for more established players with less downside and picks 11-14 on less developed players with more upside. As such, the lower picks (i.e. 11-14 vs. 6-10) offer more variability in both directions.


The following table shows that the players drafted 15th -21st overall who represent the borderline picks for the bottom 10%, bottom 25%, median, and top 25% are: Joseph Forte; Kareem Rush; Don McLean; and Erik Murdock, respectively.

  • It’s hard to imagine a less productive player than Joseph Forte, who had a total of 30 points, 17 rebounds, and 17 assists in his 25-game, 125-minute NBA career. Amazingly, he had a negative win share total, which is similar to someone getting less than 200 on a section of the SAT. With respect to the SAT, test takers get 200 points simply for filling in their names, but they lose points for guessing incorrectly.  Any student with a score under 200 would have done better not to answer any questions. Any NBA player with negative win shares (like Forte) would have done better to have never left the bench.
  • Best remembered for giving the Lakers 4 points and 15 minutes per game in the playoffs during the Lakers run to the 2004 finals, Kareem Rush only averaged six ppg during his five-year career.
  • Don MacLean had one respectable season during which he averaged 18 ppg and 6 rpg on way to earning the league’s Most Improved Player award, but otherwise was a 10 (ppg) and 4 (rpg) player for five years.
american_pie
Not Don McLean (without an “a”)
dmaclean
That’s Don MacLean (with an “a”)
  • As the borderline top quartile pick, Eric Murdock averaged 10 points, 2.5 rebounds, 4 assists, and 1.5 steals per game for over nine years. If Murdock is better than 75% of all picks taken between 16th and 21st overall, the value of such a pick is akin to a Powerball lottery ticket with the hope of getting a player like Steve Nash or John Stockton.

SUMMARY OF DRAFT PICKS SELECTED 15TH – 21ST OVERALL

Identifier

Player Win Shares Percentile

Odds of Selecting a Better Player*

Potential Bust

Joseph Forte <1 10.0% > 99.0%
Disappointment Kareem Rush 2 25.0%

> 99.0%

Average Pick

Don MacLean 9 50.0% 98.2%
Good Pick Eric Murdock 25 75.0%

81.9%

Superstar

Larry Nance 110 97.3% 14.9%
Superstar Steve Nash 129 98.3%

10.0%

Superstar

Artis Gilmore 190 99.5% 3.2%**
Superstar John Stockton 208 99.6%

2.4%**

* Odds that at least one player taken with an 15th-21st overall pick from any draft will be better than the identified player.

** Based on these percentages, the odds of selecting a similarly productive player are 1:30 for Artis Gilmore and 1:40 for John Stockton.


The following table shows that the players drafted 22nd or later who represent the borderline picks for the bottom 10%, bottom 25%, median, and top 25% are: Randolph Childress; Mamadou N’Diaye; Travis Knight; and Oliver Miller, respectively.  To start, when Oliver Miller is the best player in a group, it’s not a very impressive group. With only 7.5 ppg and 6 rpg averages during his 8-year career, Miller’s nickname of “The Big O” was aimed more at his body shape than an homage to Oscar Robinson.

om

Given the number of flame-outs at this point in the draft, there’s not much difference between a bottom 10% pick and a bottom 25% pick. For instance, Childress had 124 points over 51 games and N’Diaye had 262 points over 69 games during their respective careers. As the median pick, Travis Knight averaged 3.5 ppg and 3.0 rpg during his seven-year career. If a pick in the prior group is akin to a Powerball ticket, a 1st round pick taken 22nd or later is akin to a state lottery ticket (i.e. the odds are slightly better but payout isn’t as high).

SUMMARY OF DRAFT PICKS SELECTED 22ND OR LATER

Identifier Player Win Shares Percentile

Odds of Selecting a Better Player*

Potential Bust

Randolph Childress <1 10.0% > 99.0%
Potential Bust Mamadou N’Diaye 2 25.0%

> 99.0%

Disappointment

Travis Knight 8 50.0% 97.6%
Good Pick Oliver Miller 21 75.0%

80.3%

Star

Sam Cassell 88 97.4% 14.6%
Star Rashard Lewis 91 97.6%

13.5%

Star

Vlade Divac 96 97.9% 12.0%
Star Tony Parker** 97 99.0%

11.7%

Star

Terry Porter 110 98.5%

8.5%

* Odds that at least one player taken with a low first round pick (i.e. after #22) from any draft will be better than the identified player.

** To the extent Tony Parker has earned Superstar status, it’s because of his playoff success versus regular-season statistics.

ShareEmail to someoneTweet about this on TwitterShare on FacebookGoogle+Share on Reddit

Leave a Reply