Quote:
Originally Posted by Resolute 14
Ehh, somewhat loaded, but at the same time, accurate. Stat sites will include all the stats in special teams situations, but the analytics community as a whole routinely ignores everything that is not 5 on 5. I get some of the reasoning, notably that it equalizes for all teams, and that over the course of a season, it forms a far larger data set than 5v4, 4v4, etc. will.
Also, how does "too simple" represent a flaw? I don't see how more complicated, but with no correlation to wins and losses is better than less complicated but significant correlation. That being said, I do have to concede your point as you haven't, at any point, tied any of your data to wins and losses. So you are correct that determining what relates to success is a separate discussion.
|
You have to look at special teams, I do all the time.
But your example of winning special teams is giving one team a +1 or greater on the scoreboard to start and is too much of an advantage to be statistically significant.
It's not flawed ... it's too clearly connected to wins.
And thanks ... that's why I don't get the constant witch hunt I'm defending. I like my hockey team to give up less and generate more. It's not an equation that guarantees victories, but it's a great place to target the areas of the game that the team needs to focus on.
That's it.