# Exploitation in Men’s College Basketball

The word “exploitation” has a simple definition.  A worker is “exploited” when the wage the worker receives is less than the worker’s economic contribution to the firm. In the past I argued that there are players at top programs in men’s college basketball who are exploited.

In my latest for Time.com, I extend the story to players at much smaller programs. Exploitation Is Everywhere in Men’s College Basketball argues that players at a school like North Carolina Central University are also worth more to their school than what they are receiving to attend to the university.

The post notes that I have measured the productivity of the players on the schools I mention.  I would add, I have also measured the Wins Produced of every men’s college basketball player back to the 2001-02 season (this is more than 50,000 player seasons).  In addition, I have measured Wins Produced for both the WNBA and ABA.  One should note that the models in each of these leagues is essentially the same as the model we see for the NBA.

And the model we see for the NBA is remarkably stable over time.  You can see this quite clearly when you do a simple out-of-sample forecast.  To do this, we estimate the model for some sample of past years.  You then plug in the values for a future season (numbers from outside the sample used to estimate the model) and see if you can explain wins in this future season.  The results are that you do just about as well doing the out-of-sample forecast as you do if you just included the future season in estimating model.

What this tells us is that basketball is basketball.  This is a fairly simple game that demonstrates a remarkable level of stability over time.

Of course, individual players in basketball — who are more consistent than what we see in baseball and football — can change their performance.  But that is entirely a separate question from whether or not the Wins Produced model is stable over time and is stable in different leagues.

Perhaps I will write more about out-of-sample forecasts in the future.  I have seen people try and do this and it appears this idea is not entirely well understood.

– DJ