In July, I did a statistical study on how a draft-eligible player’s past experience affects his potential future success in the NBA. The results of the study showed that players drafted at a young age were more likely to be successful than players drafted at an older age.  Since that article was published, I’ve gotten numerous e-mails questioning the validity of the study.  Most of these were just a matter of differing opinions.  However, there are still a few lingering questions.
            The biggest concern is how I determined if a player was successful or not.  I used PER (Player Efficiency Rating, developed by ESPN’s John Hollinger), a stat that measures all the statistical aspects of a player’s game.  However, although “efficiency” is used in the name of the stat, it is a bit misleading.  Playing time does have an effect on a player’s PER.  When dealing with guys who were just drafted, playing time is a tricky issue.  Although we’d all like to believe that the best players get the most playing time, sometimes this may not be the case.  A top pick in the draft may do nothing to earn playing time other than getting paid like a top pick.  Players with higher expectations will get more playing time, and consequentially, a higher PER.
            Another concern is basketball intelligence.  Although some high school players may be more talented than most college seniors, their knowledge of the game is probably weaker, simply because of a lack of experience.  It is not unreasonable to assume that older, more experienced players are more efficient, take better shots, and make smarter plays. 
            What statistic ignores playing time and focuses purely on efficiency?  The answer is offensive rating.  Offensive rating was developed by Dean Oliver, author of Basketball on Paper.  Offensive ratings for every player in the NBA can be found at www.basketball-reference.com and numerous other web sites.  Now, before we go any further…
            Yes, offensive rating only deals with offense.  Although basketball is obviously about more than just playing offense, offensive rating may be one of the best basketball statistics out there.  Despite the many advances that stat gurus have made, no comprehensive defensive stat has been developed.  So for now, we must work with what he have. 
            Offensive rating is points produced per 100 possessions.  The league average offensive rating in 2006 was 106.  The rating includes all aspects of a player’s offensive production, including stats such as assists, offensive rebounds, and free throws.  A very complex formula is used to calculate offensive rating.  For more information, check out Basketball on Paper. 
            Besides the change in the statistic being measured, this study was performed in the same manner as the last one.  I divided all players drafted from 1995-2002 into six groups: college seniors, juniors, sophomores, freshmen, high school players, and international players.  I then separated those groups into lottery picks and non-lottery picks.  I calculated each group’s average offensive rating, standard deviation of offensive rating, and bust percentage (percentage of players who did not appear in at least 50 games in their career). 

            Right away you can see that things are going to be different with this study.  In the lottery, college seniors performed the best, with an average offensive rating a half point higher than the next closest group, international players.  High school players, despite having the highest PER, had the second lowest average offensive rating.  Overall, seniors still had the highest average offensive rating, despite carrying the burden of having more late-first round picks than the other groups.

            Similar to the previous study, high school and international players remained the most unpredictable lottery picks.  Seniors and sophomores were the most predictable.  Overall, freshmen had the lowest standard deviation, and foreign players had the highest standard deviation by a wide margin.

            These results have not changed from the last study, with one exception.  An error was found in the bust percentage of non-lottery sophomores.  Their percentage is now at 18.2%.  International players still have that very high 15.4% overall bust rate. 

Lottery picks of all ages had an average offensive rating of 103.80 with a standard deviation of 6.28.  Non-lottery picks had an average offensive rating of 100.82 with a standard deviation of 7.88.  Lottery picks had a bust rate of 0%.  Non-lottery picks had a bust rate of 10.5%. 

            This study contradicts the findings of the last study.  This crushes the validity of both, right?  Wrong.  We must return to the playing time issue.  It is much harder for a player to maintain a high offensive rating when he is carrying the offensive workload for his team.  Many of the college seniors that were drafted were not expected to carry their teams.  For every senior like Tim Duncan, there are ten efficient role-players such as Shane Battier and Kerry Kittles.  However, when you take a look at the other age groups, you get more players like Kevin Garnett, Tracy McGrady, Kobe Bryant, Dirk Nowitzki, Allen Iverson, Vince Carter, and Paul Pierce.

The point is this: the younger guys with more potential are more likely to develop into superstars that carry the workload for their teams.  The older guys are more likely to develop into solid role-players that compliment those superstars.  For a team that is struggling and looking for a new identity, it makes more sense to take a risk on an international or young college player, evidenced by the PER study.  For a team looking to add to an already talented group of players, it makes more sense to take a polished and efficient senior, evidenced by this offensive rating study.  In the end, these studies are not saying anything too unbelievable.

Information from Basketball-reference.com, NBADraft.net, and www.thedraftreview.com was used in this study.  

Copyright © 2009 Basketball-Statistics.com