Football thoughts, 2010

For the NFL playoffs last year I did a blind comparison of teams based on their win percentage compared to their strength of schedule (or opponents win percentage). That worked so well that I tried it again this year, using the same technique, here is how the teams stack up:

AFC


NFC


Stronger teams are toward the top right corner, weaker teams toward the bottom left corner, and the lines allow for a better comparison between teams. The teams this year are much closer together this year than last. So numerically how does the first round of games stack up?



The Packers show the most promise of winning with only 62% confidence (as opposed to several of last years games that were above 70%) and the Cowboys show a slight chance of moving to next week with only 54% confidence.

Again, nothing is being compared here except wins vs. strength of schedule. Things like Wes Welker's injury last week will probably have a pretty dramatic effect.

Building your own reality through creative polling

After seeing this article over at Think Progress I took a quick look at the data on Obama's approval rating throughout this year.



Note the trendline of Rasmussen stays a good 5 points below that of all other polls. The two never cross, nor do they even come close.

Statistically speaking, error in polls can never be eliminated, but it should not always err in the same direction. Somebody has a flawed polling model, either Rasmussen, or everybody else.


"Everyone is entitled to their own opinion, but not their own facts."
- Senator Daniel Patrick Moynihan