NBA Draft, Revisited

I almost named this post "Calls calls calls, calls from the public," but that wouldn't have made much sense. I must confess, I really just wanted to reference Sifl and Olly. (Do yourself a favor and watch that clip!)

I've already gotten some good feedback from readers on my NBA Draft post, so I wanted to follow-up on the ideas they presented.

The first request was for the correlation between Pomeroy rating and number of draft picks strictly for Big Ten teams. Since Pomeroy changed his rating method after 2005, the raw Pomeroy rating probably isn't the best number to use, so I went instead with Pomeroy's national ranking. Here's the raw data:

I decided I'd run the correlation in both directions - by season as well as by team. Here's how the numbers came out:

First, let's look at correlation by season. The correlation in most seasons was there, but was relatively weak - this is consistent with the generalized BCS findings. I don't see anything surprising in that chart.

The other chart is a little more interesting - it shows how much each team's success is tied to having NBA draftees on their squad. Granted, sample size is an issue here, especially for teams like Northwestern and Penn State that have only one draftee each, but it's interesting to see that Illinois shows a relatively stronger reliance on NBA draftees for their success.

Michigan State and Minnesota buck the trend - they've shown basically zero correlation between NBA draftees and team success. The most interesting example is Michigan State in 2005 - they finished as the #4 team in the nation, then lost the underrated trio of Alan Anderson, Chris Hill, and Kelvin Torbert to graduation. None were drafted.

The following season, MSU returned their Big Three of Maurice Ager, Paul Davis, and Shannon Brown, along with promising sophomore Drew Neitzel, yet stumbled to #33 in the nation. After this disappointing 2006 season, Ager/Davis/Brown were all drafted into the NBA. Then, despite this loss of talent, MSU actually climbed to #15 in the nation the next year (with zero NBA picks).

I am now realizing that this MSU example illustrates a big shortcoming of my previous analysis - it only considers the players that were drafted immediately after that season. In this example, despite being big contributors and eventual draftees, Ager, Davis, and Brown were not counted for that 2005 squad, making it look much less talented than it truly was. So, to get more clarity, let's take a look at the number of players on each squad that would eventually be drafted:

Obviously, this changes things quite a bit, and gives a better measure of how much elite talent was actually on each squad. Here's the correlations now:

Note that I've excluded the 2007 and 2008 season for the team correlations - I feel there are still guys that played in those seasons that will eventually be drafted (Raymar Morgan, for example), so the data from those seasons cannot be considered "mature" yet.

As you'd expect, this better measure of NBA-level talent gives you a better correlation with team strength. However, there's yet another way we could refine our measure. We see that the most loaded Big Ten team of the past 10 years, according to eventual NBA draft picks, was the 2003 Illinois squad. Of course, four of those five eventual picks were underclassmen. So, for an even better measure of the talent level, we should instead look at the percentage of minutes played by eventual NBA picks:

This gives us probably the best measure of how much elite talent was on each team. According to this measure, the most talented Big Ten team of the past 10 years was still Illinois in 2003, with Brian Cook, Luther Head, Deron Williams, Dee Brown, and James Augustine. The Illini also place second, third, and fourth (2005, 2004, and 2002). Clearly Illinois had an incredible four year run of talent from 2002-2005.

Let's take a look at the correlations now (again excluding 2007 and 2008 for the team correlations):

The amazing thing is that most of the teams in the conference have shown very little correlation between NBA talent and overall team strength over the past 10 years. Illinois and Indiana seem to be the most dependent on NBA talent over the past 10 years.

As always, let me know if you see anything in the data that demands further exploration. I've found this exercise to be quite interesting, but, really - is anybody else just dying for some actual basketball??