Snippit
Registered User
- Dec 5, 2012
- 16,628
- 9,959
I mean, there’s no such thing as a guarantee in statistics, that’s why we produce confidence intervals and look at values like r-squared in reference to linear regressions. So yeah, of course you could have a 30 game sample that doesn’t resemble the population, but the odds of this happening are excessively low and it kind of just seems like you’re arguing in bad faith here
Maybe I’m jumping into this prematurely but are the odds of an unrepresentative 30 game sample really as excessively low as you are making it out to be? Skinner last year for example?
This is hockey where players go through streaks and slumps, not rolling dice.