View Single Post
02-07-2013, 10:43 AM
Join Date: Mar 2006
Location: Regina, SK
I like that system too, as it has a built in accounting for the generally accepted fact that "5 th in 1937 is not the same as 5th in 1987".
As far as not removing outliers goes, as long as the results in the end "work" (I.e. a distribution of players across history who hit certain benchmark scores every year that makes sense and rises slowly over time) then it's gone. What BM67 proposed hours far down enough that it's rare one would need to care that a lemieux Or gretzky at the top is an outlier because the difference between using them and not, would be a benchmark score difference of maybe three points.
View Public Profile
Find More Posts by seventieslord