First of all, commenting that the article in question focuses too much on stats is just silly -- it's like reading something by Bill James and criticizing it for being too much about stats. As creator of the DVOA and DPAR metrics, statistical discourse is the sole reason why a site like foxsports wants to publish Schatz.
Then it is an interesting discussion of DVOA and DPAR metrics, but not of quarterbacks or team sports. To the extent that these metrics are useful, they must be predictive; modeling the past within a present season to compare the two is a trifling exercise, unless the metrics could quantify the probability/likelihood of winning. Put another way: You've got Tom Brady from 2004, and Bert Jones from 1976. You want to win a Super Bowl. Who do you take?
As for the weather issre, various performances this year have prompted Schatz to attempt to devise a way to add adjustments for weather in his stats. It should prove to be a tricky task to do with as much objectivity as possible. I'm curious to see what he'll come up with for '08.
IMO, judging from the splits between playing indoors vs. outdoors in most QBs, I imagine that the positive effect from playing half your games in a dome will turn out to be quite large.
Then Schatz would do well to mention that his current palate of variables are inaccurate, in that accounting for quarterback play in and out of weather has not been considered.
I'm not against keeping statistics, and I don't much care that Fox publishes an article about Brady's season from a statistical perspective. I have long been one of those who said, "Yep, I'm pretty sure Manning's the better quarterback, can't argue with those numbers. Yep, I'm pretty sure I want Brady as my QB, can't argue with that one number: SB rings."
That's a fancy way of saying, the stats are only useful when you make a conscious decision to work in a vacuum, away from stats that are more difficult to quantify, i.e., "intangibles" such as the influence of weather (not really an intangible) and leadership.
Through this filter, we can derive a sort of fantasy football or Madden model that tells you what should happen, yet has little or no predictive power. (What did Schatz say would happen to Brady's value in the presence of Moss and Welker, prior to the opening game of the season?)
PFnV
