View Single Post
Old
09-21-2012, 11:43 AM
  #15
TheDevilMadeMe
Registered User
 
TheDevilMadeMe's Avatar
 
Join Date: Aug 2006
Location: Brooklyn
Country: United States
Posts: 45,404
vCash: 500
Quote:
- My system gives Lacroix a lot more points in 1973, 74, and 75 thanks to forgiveness for outliers, which others like TDMM refuse to do. If I was to use the #2 scorer strictly, that 91, 87, 87 becomes 84, 81, 64, which would leave them with almost identical “best six seasons” scores (Kopitar would be behind by 21 total, or 3.5 per season) with not much after his “best six” to help Lacroix’s case. Something to keep in mind if you don’t believe in removing outliers like the hellish Orr/Esposito combo.
That's not exactly true. I think that if you are comparing to #2, you should compare to #2, except when #2 is an extreme outlier (like 1989). But outliers should be strictly and mathematically defined to remove arbitrariness.

I also prefer comparing to 5th place, rather than 2nd place, for post expansion players because it removes outliers for you.

Here is Bellows vs Satan using vs5 scores:

Bellows: 88, 73, 70, 70, 68, 67, 67, 61, 64, 51, 50, 43, 42, 38, 32
Satan: 92, 79, 77, 69, 70, 65, 64, 58, 53, 43, 39, 38, 30

Satan's "big" percentage score was 2002, when he was 14th in scoring with 72 points, but 5th place had only 79 points.

In Bellows' "big" season, he scored 99 points in 1990, which was 16 points higher than his second best season from a raw points perspective, so his high score for that season makes sense.

TheDevilMadeMe is offline   Reply With Quote