Quote:
Originally Posted by Macindoc
If you predicted that every single team would get 95 points, the average error rate would be pretty close to 10.4 points. The model is next to useless.
|
11.97 actually. Which seems close but isn’t.
I don’t think it’s meant to be useful beyond entertainment/something to talk about. It is interesting that it’s the least inaccurate model. And I think he uses the term “least inaccurate” as opposed to “most accurate” because… you know… nobody can actually predict NHL standings before the season starts with anything resembling good accuracy.