Quote:
Originally Posted by GranteedEV
Except.
It.
Did.
Bennett's numbers with Brouwer are an outlier for his aggregate numbers this season. They are SKEWING the results. Entirely. They "should not". But they "do".
Do I need to explain skew to you?
|
No, you don't. This is not skewing, this is an outlier data point that has no bearing on anything because it accounts for less than 1% of the data. A trend could not be skewed by something that accounts for that little of the overall data. If it had equal weighting to the rest of the data, yes it could skew things drastically.
Weighting matters here. If you had a bunch of work in a course and you scored 90% on assignments that added up to 99% of the course work, you wouldn't care at all that you got a 0% on an assignment that was weighted at 1%. Well in this case, it's less than 1%, so the weighting based on time on ice means that Brouwer's effect on Bennett is nil.
It's the last I'll say on this because this thread is about Jankowski. Thanks for derailing it.