Tuesday, June 21, 2011

Off the Index: Defending Allen Iverson

The 2011 NBA Draft is Thursday and the typical articles on draft busts have been popping up on the blog circuit. An article at Hoopism was very good but an article at the new Wages of Wins Network blog, Shut Up and Jam, could use some work.

Allen Iverson has been a popular target for criticism on the Wages of Wins Network because the statistical models published in the Wages of Wins illustrated that he didn’t have the winning impact the media claimed. There’s no denying that Iverson’s poor shooting percentage and turnovers created as many problems as his scoring and steals created highlights, but James Brocato made a mistake when he used wins produced to label Iverson a mistake as the first pick in the 1996 draft.

Yes, it’s a little silly to defend an 11-time All-Star and future hall-of-famer labeled a “bust” by a blogger, but it just takes a little context and the same statistical models to illustrate why Iverson was the right choice for the Philadelphia 76ers in the summer of 1996.

This article will use Win Score, a statistical model created by Professor David Berri from the Wages of Wins Journal, to measure how much a player's box score statistics contributed to their team's efficiency differential and wins. More information on these stats can be found at the following links:

Simple Models of Player Performance
Introducing PAWSmin — and a Defense of Box Score Statistics

The Philadelphia 76ers in 1996
The Philadelphia 76ers had the second-worst record in the NBA in 1996 at 18-64 (the Vancouver Grizzlies were worse at 15-67). Obviously the 76ers were bad and needed a lot of talent, but where did they need it the most? This spreadsheet lists the Win Score per 48 minutes (WS48) for the 76ers guards, forwards and centers in 1996.

The average WS48 for an NBA guard from 1994 to 2005 was 6.2. The average WS48 for a 76ers guard in 1996 was 3.7. That’s 40 percent less than average. Jerry Stackhouse (3.8 WS48), Vernon Maxwell (3.7 WS48) and Trevor Ruffin (4.2 WS48) played 85 percent of the minutes in the 76ers backcourt and they were terrible. Stackhouse was a rookie and Ruffin was in his second season, so they could improve, but Maxwell was 31 and declining.

The average WS48 for an NBA forward from 1994 to 2005 was 8.8. The average WS48 for a 76ers forward in 1996 was 10.0. Clarence Weatherspoon was an above average forward with a 12.2 WS48 that played 78 percent of the available minutes. The 76ers expected to be set in the frontcourt the next season with Derrick Coleman returning from injury. Coleman produced an average of 13.9 WS48 from 1991 to 1995 so the 76ers would have two above average forwards if he stayed healthy (see this spreadsheet for stats from Coleman’s first five seasons).

The average WS48 for an NBA center from 1994 to 2005 was 10.8. The average WS48 for a 76ers center in 1996 was 9.0. That’s 17 percent less than average, but the “true” centers on the roster only played 32 percent of the available minutes. Coach John Lucas used power forwards like Sharone Wright (9.2 WS48), Tony Massenburg (9.1 WS48) and Ed Pinckney (13.8 WS48) at center for 66 percent of the available minutes.

As illustrated by the numbers above, the glaring weakness in the 76ers’ lineup was the backcourt. Since Stackhouse was the team’s star rookie at shooting guard (and young enough to improve), the focus was going to be improving the point guard position since an aging Vernon Maxwell was not going to be the solution.

The Best College Point Guard in 1996
Who was the best point guard in the 1996 draft? This spreadsheet lists the position-adjusted Win Score per 40 minutes (PAWS40) for the first and second-team All-Americans in 1996.

Allen Iverson was the 4th-most productive All-American entering the NBA draft in 1996 (Tim Duncan did not enter the draft until 1997). He was the most productive point guard with a 12.5 PAWS40. At 6’1”, Tony Delk was going to have to play point guard in the NBA and he was more productive than Iverson in college with a 13.4 PAWS40, but he played shooting guard for the national champion Kentucky Wildcats and only averaged 2.7 assists per 40 minutes. Not exactly the solution for a team looking for a point guard to pair with Stackhouse. Iverson, on the other hand, averaged 5.7 assists per 40 minutes.

There were only two All-American point guards in 1996: Iverson (12.5 PAWS40) and Jacque Vaughn (8.6 PAWS40). It was an easy decision for the 76ers to make.

For those that would suggest the 76ers should have gone with size and chosen Player of the Year Marcus Camby to improve their below average production at center, there are two counter-arguments:

  • First, the production at point guard was much worse than the production at center; and
  • Second, Camby was less productive than Iverson in college with a 9.4 PAWS40.

By the numbers, Iverson was indeed The Answer for the 76ers in the 1996 draft.

Go to Bleacher Report and vote on who you think the 76ers should've taken with the No. 1 pick in 1996: Iverson, Camby or Nash.


  1. Another great post Mosi.

    What about the argument that they could have taken Steve Nash?

  2. Great question, Kabelo. Nash's PAWS40 was only 9.2 his senior season at Santa Clara, so AI was more productive than him, too. I didn't include Nash in the article because there have only been 2 college players taken No.1 in the last 35 years that weren't All-Americans - Derrick Rose & Michael Olowokandi.

  3. Let me make a couple of things clear. I didn't call Iverson a "bust," and my list is not based on what Philadelphia knew or should have known at the time. The reason I said Iverson turned out to be a bad #1 pick was because 1. he produced at a BELOW AVERAGE rate for his career and his best WP48 season was 0.135, and 2. 3 players in the draft produced at >0.200 for their careers and ALL 3 are still playing and still very productive. Again, the post says outright that it is using the power of hindsight, not what the 76ers knew at the time. Of all the #1 picks since 1977, Iverson had the 8th lowest career WP48.

  4. If Iverson had a better attitude he would've been one of the best players to ever play the game. Its a shame he is playing overseas and no one would take a chance with him down here. He would make a great bench player for a team.

  5. @jbrocato:

    The 2nd sentence of your post says, "Over the last 30 years, there have been a lot of busts at #1."

    I think per minute production is a bad tool for evaluating a No.1 draft pick. There's no other player on your list that's within 10 wins produced of Iverson and there are five other players that produced less estimated wins than he did. You mentioned three: Kenyon Martin, Danny Manning and Ralph Sampson. But Larry Johnson and Pervis Ellison also produced less wins than Iverson even though they had a higher Est.WP48 (see http://bit.ly/ifa9ql).

  6. @jfarbman:

    Iverson's attitude is one of the reasons I was such a big fan. It actually made me sad to hear him say that he was now willing to come off the bench in the NBA. It's something he shouldn't have to do.

  7. Did I miss something? Isn't it June 2011? How could we possibly still be talking about Allen Iverson? Wasn't his last meaningful game like 7-8 years ago?

  8. @canis39:

    Yes, you missed something: the 1st two paragraphs of the article.