DAChampion
Registered User
- May 28, 2011
- 30,203
- 21,650
But then you wouldnt take into account that Josh is an outlier age wise in his draft year. Youd have to go by decimal years and months.
Then you'd probably need to run a windsorized mean and even so, is it actually accurate if you do so?
Then how do you modelize for AHL and NHL team depth? Or for Euro players that had to adjust to NHL ice?
You would get a very rudimentary statistic that I am not sure would actually reflect reality. I understand were not completing a peer reviewed statistical analysis here but the quality of such a statistic is heavily influenced by factors well beyond our reach.
That would actually worsen the fit. You don't want to double count variables or overfit in other ways.