Kaapo Bustfreniere
Registered User
- Jun 21, 2019
- 458
- 709
Again totally missing the point of my original point.If you argue with statistics, more data is in most cases more predictive, than excluding a certain sample. Should we exclude games against Philly, because those mini series don't happen in a normal NHL season?
But since you want to argue about statistics:you always want to have data that are as replicable as possible. If we want to apply that on this example, then if a hockey player is streaky or often suffers from say neck injuries that slow him down, that is something that will likely keep on happening. So if you include data from both his hot and cold streaks and from games where he suffered from repeated injuries, that is fine. Bacause that would be a representative sample - which you always want. 3 points in 15 games at the beginning of the season is a huge statistical outlier for Zibanejad´s production - even more so when you look at the last 3 ears since he really broke out. And suffering from a postcovid syndrome that he seems to have recovered from is something that will most likley never happen again. As such, it does NOT present a representative sample and can be excluded.