Illustration by Michael Walchalk

Stat Stories: Lucky Stars

Quantifying the NBA Draft’s Most Successful Picks, Teams and GMs

The NBA Draft is a crapshoot with the odds of landing a franchise-altering player down to that of a simple coin flip. Yet, teams have gotten lucky landing a superstar in the lottery while others continually find talent regardless of where they’re picking. But with a multitude of variables involved, determining successful picks, teams and general managers has become increasingly difficult, so Chad Shanks and Justin Kubatko explain how a custom calculation that weighs a player’s draft position against his career output can quantifiably determine who’s been the best at finding value on draft night.

Listen to the embedded audio and browse through the StatMuse stats mentioned in the episode (data accurate as of the date of publication).

Justin Kubatko’s “Ku-Draft-Ko” System*

Back in 2013 I did some work for ESPN Insider where I evaluated the careers of NBA lottery picks, but one thing that always bugged me about that analysis was the inherent assumption that all drafts are created equal. In other words, the expected career value of the first overall pick was always the same regardless of the talent available. This is false, of course, so I wanted to come up with a way to account for this quirk.

In order to do this I first had to come up with a way to assign a career value to every player. I decided to use a weighting scheme popularized by Doug Drinen, creator of Pro Football Reference. Here’s Doug’s description of the method:

My opinion is that most people mentally rank players by counting all the players’ seasons, but weighting their best seasons more. In order to mimic that, I’ve defined each player’s approximate career “value” to be:

100 percent of his best season, plus 95 percent of his second best season, plus 90 percent of his third best season, plus…

So, for two players with the same career value, the one with the higher peak will be rated a little higher. And junk seasons at the end of a player’s career count for almost nothing.

Using Drinen’s method as my blueprint, here’s what I did:

1. A player’s season value is equal to his regular season wins above replacement.
2. The player’s season values are ordered from best (highest) to worst (lowest).
3. The player’s career value is equal to 100 percent of his best season, plus 95 percent of his second-best season, plus 90 percent of his third-best season, etc.

I calculated the career value for every first and second round draft pick from 1985 (the first year the lottery was used) through 2012. It was assumed that the first overall pick would produce the highest career value, the second overall pick would produce the second-highest career value, etc. Then within each draft year I compared each player’s career value to the career value of every other player chosen in the first two rounds of that draft. The expected and observed ranks are converted into a draft score using the following formula:

draft_score = ln(expected_rank / observed_rank)

For example, a player drafted with the 10th overall pick who produces the second-most career value in his draft would receive a score of:

draft_score = ln(10 / 2) = ln(5) = 1.61

There are, of course, numerous ways one could do this, but I like this method for several reasons:

1. The players aren’t saddled with unreasonable expectations simply based on their draft slot. In other words, a high lottery pick in a weak draft doesn’t have to be a star in order to receive a decent score.
2. Players who finish higher than their expected rank will have a positive draft score, players who finish lower than their expected rank will have a negative draft score, and players who finish at their expected rank will have a draft score of zero.
3. Getting the player with the highest career value with the 11th overall pick (+2.40) is judged to be much more valuable than getting the player with the 15th-highest career value with the 25th overall pick (+0.51) even though each player finished 10 spots higher than expected. I think this makes intuitive sense.

As an example, here are the draft scores for the 13 lottery picks in the 1998 NBA Draft:

1. Michael Olowokandi (LAC), –4.03
2. Mike Bibby (VAN), –1.10
3. Raef LaFrentz (DEN), –1.10
4. Antwan Jamison (GSW), 0.00
5. Vince Carter (TOR), +0.51
6. Robert Traylor (DAL), –1.34
7. Jason Williams (SAC), –0.69
8. Larry Hughes (PHI), –0.41
9. Dirk Nowitzki (MIL), +2.20
10. Paul Pierce (BOS), +1.61
11. Bonzi Wells (DET), –0.17
12. Michael Doleac (ORL), –0.69
13. Keon Clark (ORL), –0.38

This method makes it possible to answer any number of draft-related questions, so the rest of this post will be presented in a question-and-answer format. As you read, please keep in mind that the results below are for the 1985 through 2012 NBA Drafts only, not every draft in history.

Also, please note that this really isn’t a system for comparing one player to another. It’s more of way to compare teams or executives over a long period of time. When you’re looking at 1,572 picks over a 28-year period, you really can’t assign a subjective score to each pick.

Who were the biggest busts?

1. Michael Olowokandi (1998, 1st pick), –4.03
2. Greg Oden (2007, 1st pick), –3.22
3. Len Bias (1986, 2nd pick), –3.16
4. Kwame Brown (2001, 1st pick), –2.94
5. Hasheem Thabeet (2009, 2nd pick), –2.92

What is Michael Olowokandi’s offensive rating by season?

Bias, the 1985 and 1986 ACC Player of the Year from Maryland, sadly never got the chance to play for the Boston Celtics, as he died of a drug overdose just two days after the 1986 NBA Draft.

Who were the biggest steals?

1. Jeff Hornacek (1986, 46th), +3.83
2. Michael Redd (2000, 43rd), +3.76
3. Isaiah Thomas (2011, 60th), +3.40
4. Marc Gasol (2007, 48th), +3.18
5. Paul Millsap (2006, 47th), +3.16

What is Jeff Hornacek’s offensive rating by season?

The 1986 NBA Draft was an odd one, as six of the top nine players in career value were selected with the 24th pick or later.

Given the season he just had, it’s hard to believe that Isaiah Thomas was the last pick in the 2011 NBA Draft.

How many players drafted first overall produced the highest career value in their draft?

1. David Robinson (1987)
2. Shaquille O’Neal (1992)
3. Chris Webber (1993)
4. Tim Duncan (1997)
5. LeBron James (2003)
6. Dwight Howard (2004)
7. John Wall (2010)
8. Anthony Davis (2012)

A summary of the 28 first overall picks from 1985 through 2012:

• Eight (28.6 percent) finished with the highest career value in their draft.
• Three (10.7 percent) finished with the second-highest career value in their draft.
• Three (10.7 percent) finished with the third-highest career value in their draft.
• Three (10.7 percent) finished with the fourth-highest career value in their draft.
• Four (14.3 percent) finished outside the top 10 in career value in their draft class (Pervis Ellison, Michael Olowokandi, Kwame Brown, and Greg Oden).

Which teams were the most successful?

By total score:

1. Phoenix Suns (64 picks), +21.34
2. San Antonio Spurs (53 picks), + 14.85
3. Los Angeles Lakers (46 picks), +14.28
4. Oklahoma City Thunder (68 picks), +13.76
5. Utah Jazz (54 picks), +12.01

By average score:

1. Phoenix Suns (64 picks), +0.33
2. Los Angeles Lakers (46 picks), +0.31
3. San Antonio Spurs (53 picks), +0.28
4. Utah Jazz (54 picks), +0.22
5. Oklahoma City Thunder (68 picks), +0.20

The Suns have selected three players who produced the most career value in their respective draft class, all with the 9th pick or later:

• Jeff Hornacek (46th pick in the 1986 draft)
• Shawn Marion (9th in 1999)
• A’mare Stoudemire (9th in 2002)

The Suns also drafted Michael Finley with the 21st pick in the 1995 NBA Draft (3rd-highest career value in his draft class), Dan Majerle 14th in 1988 (4th), Steve Nash 15th in 1996 (3rd), and Cedric Ceballos 48th in 1990 (6th).
Now is probably a good time to note that the team that drafted the player gets full credit for the pick even if the player was traded to another team. There are certainly ways one could adjust for this, but for the most part I think it’s a wash over such a long period of time. For example, even though the Lakers don’t get credit for drafting Kobe Bryant, they do get credit for drafting Marc Gasol, who never played a minute for the team (and was actually a bigger steal than Bryant).

Which teams were the least successful?

By total score:

1. L.A. Clippers (56 picks), –25.52
2. Minnesota Timberwolves (57), –11.85
3. Washington Wizards (57), –11.66
4. Brooklyn Nets (48), –10.19
5. Atlanta Hawks (64), –9.99

By average score:

1. L.A. Clippers (56), –0.46
2. Toronto Raptors (30), –0.30
3. Memphis Grizzlies (36), –0.27
4. Charlotte Hornets (42), –0.23
5. Brooklyn Nets (48), –0.21

Which team has the lowest winning percentage since 1985?

The Clippers selected 11 players in the top five from 1985 to 2012 — two more than any other team — and incredibly every single one of the those players has produced a negative draft score. Their notable misses include:

• Benoit Benjamin (3rd pick in 1985, 15th in career value)
• Reggie Williams (4th in 1987, 13th)
• Danny Ferry (2nd in 1989, 12th)
• Michael Olowokandi (1st in 1998, 56th)
• Darius Miles (3rd in 2000, 22nd)
• Shaun Livingston (4th in 2004, 21st)

Which executives were the most successful?

By total score:

1. Bryan Colangelo (30 picks), +14.01
2. Jerry West (30), +10.39
3. Gregg Popovich (15), +7.72
4. Geoff Petrie (39), +7.68
5. Don Nelson (38), +7.33

By average score (minimum 10 picks):

1. Gregg Popovich (15 picks), +0.51
2. Bryan Colangelo (30), + 0.47
3. Larry Bird (15), +0.45
4. Sam Presti (17), + 0.35
5. R.C. Buford (20), +0.35

Since the Suns were on top of the list of most successful teams, it’s not a surprise to see Bryan Colangelo here. Colangelo was able to get a number of steals in the draft:

• 1994 — Wesley Person (23rd overall pick, 9th in career value)
• 1995 — Michael Finley (21st, 3rd)
• 1996 — Steve Nash (15th, 3rd)
• 1997 — Stephen Jackson (42nd, 7th)
• 1999 — Shawn Marion (9th, 1st)
• 2002 — A’mare Stoudemire (9th, 1st)
• 2005 — Marcin Gortat (57th, 5th)

Which team has the highest offensive rating since 1994?

Colangelo’s biggest miss, by far, came with the Toronto Raptors in 2006, when he selected Andrea Bargnani with the first overall pick. Bargnani currently ranks 10th in his draft class in career value.

Andrea Bargnani career shot chart with Toronto

Which executives were the least successful?

By total score:

1. Elgin Baylor (49 picks), –22.97
2. Jerry Krause (48), –8.70
3. David Kahn (14), –6.71
4. John Nash (27), –5.80
5. Stu Jackson (12), –6.10

By average score (minimum 10 picks):

1. Stu Jackson (12 picks), –0.51
2. David Kahn (14), –0.48
3. Elgin Baylor (49), –0.47
4. Willis Reed (10), –0.38
5. Glen Grunwald (13), –0.35

You might be surprised to see Jerry Krause’s name appear in the first table, but his draft record is not particularly strong. Krause picked up some steals for the Chicago Bulls early in his tenure:

• 1987 — Horace Grant (10th overall pick, 5th in career value)
• 1989 — B.J. Armstrong (18th, 10th)
• 1990 — Toni Kukoc (29th, 3rd)

But some of his others picks from the 1980s were far from stellar:

• 1985 — Keith Lee (11th overall pick, 26th in career value)
• 1986 — Brad Sellers (9th, 25th)
• 1989 — Stacey King (6th, 20th)

And his first round picks from 2000 to 2002 were abysmal:

• 2000 — Marcus Fizer (4th overall pick, 28th in career value)
• 2000 — Chris Mihm (7th, 18th)
• 2000 — Dalibor Bagaric (24th, 49th)
• 2001 — Eddy Curry (4th, 15th)
• 2002 — Jay Williams (2nd, 30th)

Who had the lowest offensive rating as a rookie in 2000–01 (minimum 10 games)?

*Justin Kubatko in no way endorses the “Ku-Draft-K0” moniker and actually wishes Chad would stop using that ridiculous name**.

**Chad will continue to refer to it as the “Ku-Draft-Ko” system despite Justin’s objections.

Sources/Further Reading

Why Top NBA Draft Picks So Often Disappoint

The Savior Fallacy: Over-Betting on Star Players in Sports and Business

Additional Stat Stories Episodes

Stat Stories: Episode III — Revenge of the Whiffs

Stat Stories: Episode 26 — Tears of a Clown

Stat Stories: Episode 27 — Imbalance of Power

Stat Stories: Episode 28 — Teflon LeBron