A colleague and I are working on a predictive analytic model for hockey. Why? Because there is data available, and it seems a nice extension and a fun thing to do. And I really like working with numbers and analyzing data.
Looking at several years of hockey win percentage and goal differentials we find the following. Seems pretty clear that there is fairly tight relationship between the goal differential and winning. Not a surprise really, but nice to confirm that what we observe for other sports also holds true in hockey. We will update as we get more results.