View Single Post
Old 11-17-2014, 05:20 PM   #175
Resolute 14
In the Sin Bin
 
Resolute 14's Avatar
 
Join Date: Jan 2003
Exp:
Default

Quote:
Originally Posted by Street Pharmacist View Post
You're wrong, though. I suppose it depends on your qualifier on "outlying", but almost universally a PDO that is significantly higher than 1 will regress to the mean. That's not fallacy, that's fact.

For example. Take any PDO over 1.02 for a season. Now add the season before it. Guess what, whatever sample you chose will now be lower. The top PDO and bottom PDO in the league all get smaller as the sample size gets bigger.

If you are calling a PDO of over 1.02 an outlyer, then yes, for one season that's possible. Colorado did it last year. How did that work out?
I think you are both right, actually.

As the sample gets larger, PDO almost always continues to regress to 1.

However, I would agree that it is fair to say PDO itself is not predictive. Certainly in the arguments I made in response to _Q_'s posts, my belief that certain teams will see a rise or drop in PDO is based off other stats - primarily Corsi and Fenwick, zone starts/finishes, faceoff locations. Basically, where the puck seems to spend much of its time. Those are the ones that are predicting a change in success.
Resolute 14 is offline   Reply With Quote
The Following User Says Thank You to Resolute 14 For This Useful Post: