View Single Post
02-07-2011, 03:13 PM
Hawkey Town 18
Hawkey Town 18's Avatar
Join Date: Jun 2009
Location: Chicago, IL
Posts: 6,153
vCash: 500
Originally Posted by seventieslord View Post
Don't worry. Very few people here want to strictly use points finishes anymore. I think that for the majority of post-expansion to post-expansion, and pre-expansion to pre-expansion comparisons, they still work very well as a starting point and usually get the point across. If you're trying to compare a pre-expansion player to a post-expansion player then bringing out the percentages is the fairest way to go. At the same time, adhering too close to the percentages will see you overrate a modern player. If a 15th-place guy in one season today out-percentages a 5th-place guy from a season in the 50s, I still want the guy in the 50s because there's something to be said for being among the top-5 in the best league on earth.

If peer comparison is the best way to rank players, isn't this a much more accurate picture of a players peers?

Yes, it appears that the difference from 5th-10th is similar to the difference from 25th-50th. But why are 5th and 25th the benchmarks? there's no way 5th in 1960 is equivalent to just 25th today, so naturally the 50th-place player today will look equal to the 10th from 1960, because they are being compared to very different things!
I chose to use 5th rather randomly. I didn't want to use the first few spots because those numbers can be skewed by the freaks (just like any statistical extreme really). 25th is the equivalent because in a 30 team league there are 5x as many 1st liner players than a 6 team league.

You say 5th and 25th are not equivalent, but in terms of where that player ranks amongst first line players in the NHL they are equivalent. The 5th ranked player is in the 72nd percentile among 18 first line forwards, and the 25th ranked is in the same spot among 90 first line forwards. When comparing to their peers, they are in the exact same spot.

Hawkey Town 18 is offline