Page 2 of 2
Re: Statman's 2014-15 WAR at hoopsnerd.com
Posted: Thu Dec 11, 2014 8:34 am
by Crow
Not shy about work. Use the dialogue to your advantage.
Re: Statman's 2014-15 WAR at hoopsnerd.com
Posted: Fri Dec 26, 2014 9:31 pm
by ampersand5
I can't remember the exact number, but in the seminole paper "pot calling the kettle black", Lewin and and Rosenbaum demonstrated that minutes per game was a much higher predictor of wins than almost all other advanced metrics.
As playing time is an important component of your metric, it is not entirely surprising that it would have great predictive power.
This seems as if it would remove a lot of the surprising results of other metrics and give less "weird" numbers.