View Single Post
12-11-2012, 05:07 PM
Student Of The Game
seventieslord's Avatar
Join Date: Mar 2006
Location: Regina, SK
Country: Canada
Posts: 30,646
vCash: 500
Originally Posted by TheDevilMadeMe View Post
I think this would work well with goalies in tandem situations, where the starter and backup both play a significant number of different teams in different situations. But it gets tough when the starter plays 70 games and the backup 12, especially since the backup's games are not going to be randomly distributed. Some coaches will play backups on the 2nd half of a back to back (which will hurt the backup's numbers). Other coaches will only play the backups against weaker teams (which will help the backup's numbers relative to a starter).
Originally Posted by Taco MacArthur View Post
Sample size issues will certainly be a concern.

I see two dimensions of QoC here - do goaltenders play against stronger or weaker competition than is reflected in save percentage, and then (once that is accounted for), do goaltenders pay better or worse than their teammates?

It would be hard to extract the signal from the noise (although perhaps not impossible).
I should have mentioned, it would be weighted. So the seasons that are insignificant sample sizes would not have a large overall impact on the goalie's career result.

As far as backups getting easier or harder starts from team to team... I can't think of a way to account for that.

seventieslord is offline   Reply With Quote