I feel like a broken record. I know Rutgers was keying on the run and giving us the pass. That doesn't change that I didn't see the results I needed to see that something had changed. Null hypothesis is that things are the same until shown otherwise.
THIS!
The other night I was Dilaudid-sext-ranting to Coker about Bayesian analysis and why you have to take all years of Petersen as meaningful prices.
He deserves a medal.
That should be 'priors' not 'prices'. Fucking AutoCorrect.
It's hard.
But yeah, as my class poaster friend @Mad_Son is saying the idea is this:
The important premise is we are predicting something that changes, but is not random.
Given that the pattern of data is not random, data already gathered can help you make more accurate (not 100% accurate, just MORE accurate) predictions about the future.
So, we just look at what we've gotten out of Pete so far... let's make it simple and not even use stats, just grades.
2014 Offensive Grade: C-
2014 Defensive Grade: B+
2015: Offensive Grade: D
2015 Defensive Grade: A-
What are the grades for 2016 likely to be?
And then, even more than that... Given our prior knowledge, what new information would we view as satisfactory evidence that a new pattern was emerging?