Roland Beech is back with more fun statistical stuff to read - a 7 part series on NBA player development. The first two parts are up so far, Expected Performance by Draft Pick Number and Best/Worst Drafting Teams.
Basically, in Part I he used player stats for drafted players over the last 20 years to calculate the expected performance by draft position. As he acknowledges, he uses a pretty simplistic formula because of the data available - it's not intended as an analysis of specific individual players, but to determine in aggregate how teams have done with drafting and development. There are issues with the data he uses (career averages - rating = P + A +R), but he pretty clearly points out the weaknesses. Of particular significance for the Bulls, more recent draft picks are less likely to have reached their peak yet and are thus more likely to rank negatively when compared to career average expectations. He also provides the percentage of players at each draft position who he classifies as star/solid/role player/deep bench/bust - but again, because the data used are pretty simplistic this won't necessarily correlate with our general idea of who is a star vs. solid or role player.
Part II looks at each team over the last 20 years, calculating their draft picks actual performance in comparison to expected performance. Milwaukie ranks first, with a differential of +1.8, and New York last, with a -1.5. Chicago is roughly in the middle (19th) with a differential of -0.3. The remaining parts look to be really interesting, too - titles are posted, but the articles aren't available yet.
With the rumors flying about Pax and all the discussion over how his career rates, these articles made me wonder how his drafting would stack up under this criteria. It seems from what I've read around here that a lot of people think he's done a very good job at drafting, or at least that he did in the early years then wasn't so strong starting with 2006. Using the formula from the 82games.com articles, I came up with an overall differential for Pax of -2.3, which would rank him as a horrible drafter. However, the differentials for Joakim, Thabo, and especially Ty are pretty bad and bring the overall score down, which also supports the post-2006 criticisms. Keep in mind, however, the caveat that young players who were high draft picks are more likely to show poorly under this formula. In the next couple of years, with Tyrus and Joakim playing more minutes (thus higher per game averages we would hope) and better (again, we would hope), Pax is likely to have a significantly better ranking. Also, I left out Omer Asik - he was drafted with the understanding he wouldn't play for a few years, and his -5.5 would really skew the results.
If you look at the older guys, pre-2006 (whose career averages are pretty consistent with current play), Pax actually does show as a very good drafter, with a differential of +17.2. Also, if you totally cheat and just use Tyrus, Joakim, and Thabo's current year stats, Pax's rating is a +1.4. (Yeah, I realize it's horrible to establish a methodology and then cheat because you don't like the results. What's worse is that my motivation to cheat is so Tyrus doesn't look so bad, even though this isn't intended and would be a lousy assessment of an individual player.)
Here are the individual player differentials - but keep in mind this isn't designed to be representative of them as individual players, just as a rough system for evaluating the drafter. (I'm just waiting for the posts flipping out that Kirk and Lu have "actual" ranks so close to Ben's)
Kirk (7) - 23.9 actual - 17.8 expected - +6.1
Mario Austin (36) -0 actual - 5.5 expected - -5.5
Matt Bonner (45) - 10.8 actual - 7.8 expected - +3.0
Tommy Smith (53) - 0 actual - 2.6 expected - -2.6
Ben (3) - 24.3 actual - 23.8 expected - +0.5
Luol (7) - 23.9 actual - 17.8 expected - +6.1
Duhon (38) - 15.3 actual - 5.7 expected - +9.6
Tyrus (4) - 12.5 actual - 22.2 expected - -9.7
Thabo (13) - 9.3 actual - 15.4 expected - -6.1
Joakim (9) - 13.2 actual - 16.6 expected - -3.4
Aaron (49) - 8.4 actual - 5.1 expected - +3.3
PeeOn (51) - 0 actual - 3.5 expected - -3.5
Derrick (1) - 26.9 actual - 27.0 expected - -0.1
Omer Asik (36) - 0 actual -5.5 expected - -5.5
Also, I wasn't really sure how best to handle some of the guys - I mean, it's fair to treat PeeOn as a DNP, but doesn't seem quite so fair to treat Asik that way when we knew he wouldn't be playing here for a couple of years. In the end I just included everyone who was credited to the Bulls as a draft pick, no matter how long they remained on the team, with the huge exception of Asik. I'm not really sold on that being the best methodology by any means - but this was the point where I got really tired of this. ;)