Thread: Bryan Trottier
View Single Post
08-28-2010, 08:53 AM
Registered User
Canadiens1958's Avatar
Join Date: Nov 2007
Posts: 15,287
vCash: 500
The Fractional Child

Originally Posted by BraveCanadian View Post
Anyone who truly understands statistics also knows that averages don't apply very well to outliers. Even using its simple method you can immediately see how far out in front the Canadiens were.. even in comparison to some other dynasty teams.

It was easier for the Canadiens to score by a large margin over the average team of the league that year too. They weren't only better at defence, they were better at everything by a large margin.

You guys need to look at even the simple rating system that hockey-reference uses.

I can fully accept the validity of applying adjusted stats to players and teams that are close to the averages of the years that are being compared. They are reasonably accurate for that but the farther a team/player in any given year is away is from the average of that year, the worse and worse adjusted stats can explain what they did. It implicity assumes that the level of competition in the league during each year and every year is exactly the same.

And that simply is very far from the truth in some cases.
Averages reflect a collective situation but they do not reflect reality. Collectively a typical family may be represented as having 2.3 children in a specific demographic. But the fractional child never actually exists. Proceeding as if the fractional child actually exists quickly derails the process.

Canadiens1958 is online now   Reply With Quote