Quote:
Originally Posted by cdcox
jAZ -- The analysis that Brian Burke does is significantly more sophisticated than the way you've looked at it. I've asked him to run the numbers of Bob Sanders. I agree that it is a good case to consider.
Third Eye -- WPA is calculated based on play-by-play data. So you have hundreds of raw data points for with and without TP. The difference between with and without TP is large. I think it is a mistake to say that the study has no statistical significance.
|
There are simply too many omitted variables IMO. I honestly wouldn't have the slightest clue how to go about testing his system, but I'm fairly confident that if it were transformed into a linear regression or something similar the results wouldn't be conclusive. I could do it, I've got the software, but I don't feel like spending my day off getting it together.